How the Brain Processes Images
Subject: How the Brain Processes Images
57 chapters
1. A Reference Curriculum on the Biological Vision Stack: Architecture, Signals, and the Evidence That Pins It Down
[Verse 1]
In the retina's depths, photoreceptors dance
Rods and cones convert photons to a neural stance
Bipolar cells relay the signals through the chain
While horizontal cells spread contrast across the plane
Ganglion cells fire their coded message true
Creating center-surround fields in patterns we pursue
[Chorus]
From photons to perception, layer by layer we climb
Retina to cortex, processing visual time
Edge detectors firing, orientation maps align
Complex cells build features from simple design
The biological vision stack, refined through evolution's grind
From light waves to meaning in the visual mind
[Verse 2]
Lateral geniculate nucleus sorts the visual stream
Magnocellular pathways track motion's fleeting dream
Parvocellular channels carry color and fine detail
Through optic radiations that never seem to fail
Primary visual cortex with columns neat and straight
Simple cells detect edges at angles we calculate
[Chorus]
From photons to perception, layer by layer we climb
Retina to cortex, processing visual time
Edge detectors firing, orientation maps align
Complex cells build features from simple design
The biological vision stack, refined through evolution's grind
From light waves to meaning in the visual mind
[Bridge]
Hubel and Wiesel mapped the cortical terrain
Found receptive fields that drove the visual train
What pathway processes where and how things move
Ventral stream determines what we need to prove
Hierarchical processing builds up complexity
From bars and edges to full object reality
[Verse 3]
Area seventeen starts the cortical cascade
Simple cells respond to oriented lines displayed
Complex cells pool inputs from their simple kin
End-stopped cells detect where corners begin
Higher areas integrate across the visual field
Building representations that object recognition yield
[Verse 4]
Temporal cortex assembles features into forms
Inferior temporal regions weather recognition storms
Face patches specialized for human countenance
While other modules handle objects at a glance
Feedback connections refine the forward flow
Top-down predictions help us truly see and know
[Chorus]
From photons to perception, layer by layer we climb
Retina to cortex, processing visual time
Edge detectors firing, orientation maps align
Complex cells build features from simple design
The biological vision stack, refined through evolution's grind
From light waves to meaning in the visual mind
[Outro]
Architecture proven through electrode's faithful trace
Signals mapped and measured in this visual space
Evidence accumulated through decades of research
The vision stack decoded through scientific search
2. The stack at a glance
[Verse 1]
Light comes dancing through your pupil door
Hits the retina like waves on shore
Photons convert to electric fire
Neural computation starts to wire
Before a signal leaves the eye
Twenty streams begin to fly
Parallel channels, partly same
Playing the visual processing game
[Chorus]
From retina to LGN relay
Ninety percent will find their way
To V1 where the paths divide
Ventral stream for what's inside
Dorsal stream for where and how
The visual stack is working now
Two highways in your cortex maze
Processing all your visual days
[Verse 2]
Lateral geniculate holds the gate
Thalamus checkpoint, information freight
Primary visual cortex takes the lead
V1 neurons plant the seeing seed
Then the signal splits in two
Different jobs for each crew
Retinotopic maps unfold
Dozen areas, story told
[Chorus]
From retina to LGN relay
Ninety percent will find their way
To V1 where the paths divide
Ventral stream for what's inside
Dorsal stream for where and how
The visual stack is working now
Two highways in your cortex maze
Processing all your visual days
[Bridge]
Ventral sliding toward the temple zone
Inferotemporal, identity's throne
What am I seeing, face or tree
Object recognition, mystery free
Dorsal climbing parietal heights
Motion and space, action's sights
Where is it moving, how to reach
Two streams with lessons each will teach
[Verse 3]
That's the cartoon, clean and neat
But reality's more bittersweet
Hidden complexity runs deep
More than simple paths we keep
Visual guidance needs both sides
Cross-talk where the signal rides
The stack's more tangled than we thought
Beautiful chaos, finely wrought
[Verse 4]
Evolution built this seeing machine
Million years of pressure keen
From ancient light-sensing spots
To cortical processing lots
Every layer serves a role
In constructing vision's goal
From photon to perception's birth
The miracle of sight on Earth
[Final Chorus]
From retina to LGN relay
Multiple pathways light the way
Beyond V1 where streams divide
Ventral, dorsal, side by side
Processing visual information flow
The stack has secrets yet to show
In your cortex, ancient and new
The world takes shape in front of you
3. Five facts that change how you should think about it
[Verse 1]
Your eye's not a camera clicking shots in the night
The retina's a neural net, processing light
Extracting contrast, motion, colors that oppose
Direction selectivity, before the data flows
Through optic nerves at one megabit per beat
Heavy compression makes the signal complete
[Chorus]
Five facts will shake your vision foundation
Retina processes, not just sensation
Pathways are many, feedback's the king
Areas blur and timing's the thing
That proves the hierarchy we're navigating
Brain's image processing, more fascinating
[Verse 2]
More than two streams flow from your seeing machine
Not just what and where, but pathways unseen
Ten ganglion types project to different zones
Cortical areas connected like nodes
Hundreds of links form a tangled web graph
Not a simple tree, that's just teaching math
[Chorus]
Five facts will shake your vision foundation
Retina processes, not just sensation
Pathways are many, feedback's the king
Areas blur and timing's the thing
That proves the hierarchy we're navigating
Brain's image processing, more fascinating
[Bridge]
Felleman and Van Essen mapped the wiring maze
Feedback connections dominate the neural ways
Lateral and descending, outnumbering the feed
Forward signals settle in a looping read
Perception's not a sweep, it's a settling dance
Back and forth until the circuits find their stance
[Verse 3]
V4 boundaries shift like sand between your hands
Retinotopy, cytoarchitecture, connection strands
Functional responses, criteria that clash
Fuzzy borders, species differences flash
Human homologs still spark debates that rage
Macaque and human brains on different page
[Chorus]
Five facts will shake your vision foundation
Retina processes, not just sensation
Pathways are many, feedback's the king
Areas blur and timing's the thing
That proves the hierarchy we're navigating
Brain's image processing, more fascinating
[Verse 4]
But timing tells the truth with millisecond precision
V1 to V2 to V4, clean decision
IT comes last in this latency cascade
Response delays reveal how signals parade
Topological sorting of the neural stack
Timing's the compass keeping theory on track
[Outro]
So remember these five when you study the brain
Vision's more complex than textbooks explain
Neural networks, feedback loops, and timing's reign
That's how your brain processes the visual domain
4. The framing to carry forward
[Verse 1]
Your eyes ain't cameras clicking frames in sequence neat
Vision's more like jazz musicians trading riffs they meet
Lower neurons catch the edges, corners sharp and small
While higher brain cells dream of faces, shapes that never fall
It's hierarchical prediction on a graph that bends and flows
Not some factory line assembly, more like how a river knows
[Chorus]
Feedforward carries evidence ascending like the blues
Feedback drops predictions down with attentional cues
Small receptive fields below catch details fleeting fast
Large receptive fields above hold memories that last
It's the interaction where the magic happens true
Recurrent graph computing everything you view
[Verse 2]
Down in V1 they're encoding spatial frequency high
Local patches, rapid changes, movements of the eye
Up in temporal cortex dwelling sit the abstract kings
Invariant to transformation, recognizing constant things
Your grandma's face stays grandma's face from near or far away
'Cause higher levels learned the rules that never seem to sway
[Chorus]
Feedforward carries evidence ascending like the blues
Feedback drops predictions down with attentional cues
Small receptive fields below catch details fleeting fast
Large receptive fields above hold memories that last
It's the interaction where the magic happens true
Recurrent graph computing everything you view
[Bridge]
When you spot a friend approaching through the crowded street
Bottom-up meets top-down signals where the pathways meet
Prediction error drives the learning, updates what you know
Attention amplifies the signal, makes important features glow
This ain't no simple bottom-up conveyor belt design
It's bidirectional conversation, your predictive mind
[Verse 3]
Fast-changing information travels upward seeking home
Slow invariant properties drift down like honeycomb
Context shapes perception through those feedback highways wide
While feedforward pathways carry sensory data's ride
The cortical hierarchy builds reality complete
Through recurrent computations, making vision so elite
[Chorus]
Feedforward carries evidence ascending like the blues
Feedback drops predictions down with attentional cues
Small receptive fields below catch details fleeting fast
Large receptive fields above hold memories that last
It's the interaction where the magic happens true
Recurrent graph computing everything you view
[Outro]
So next time that you're seeing, remember what's at play
Hierarchical prediction guides you through your day
Not a camera, not a scanner, but a graph that thinks and learns
Through feedforward, feedback cycles, how your vision discerns
5. Physical layout (back to front, light travels front to back)
[Verse 1]
Light comes knocking at your eye's front door
But the story's twisted, there's so much more
Passes through the layers, crystal clear
Till it hits the back wall, photoreceptors here
Counterintuitive design, nature's strange plan
Signal flows backward then forward again
[Chorus]
Back to front, that's where light must go
Through transparent layers, nice and slow
Front to back, signals make their way
Retina's wired in a backwards display
Back to front, front to back
Neural signals on a zigzag track
[Verse 2]
Rods and cones are waiting at the rear
One hundred twenty million rods appear
Six million cones in three different types
Short and medium, long wavelength stripes
Blue and green and red receptors shine
Packed together in the foveal line
[Chorus]
Back to front, that's where light must go
Through transparent layers, nice and slow
Front to back, signals make their way
Retina's wired in a backwards display
Back to front, front to back
Neural signals on a zigzag track
[Verse 3]
Rods work best when darkness fills the scene
Saturate in daylight, can't be seen
Periphery patrol, but fovea's bare
Scotopic vision everywhere
Cones need brightness, color is their game
Photopic processing stakes their claim
[Bridge]
Photon hits retinal, twists the key
G-protein cascade, transducin spree
Phosphodiesterase breaks cGMP down
Sodium channels close without a sound
Hyperpolarize when light arrives
Glutamate drops, that's how vision thrives
[Verse 4]
Horizontal cells spread signals wide
Lateral inhibition, side to side
Center-surround fields, contrast they make
Ambient adaptation for vision's sake
Bipolar cells split the signal stream
ON and OFF channels, the retinal scheme
[Verse 5]
Ganglion cells process the final code
Magno, parvo, konio on the road
Large cells carry motion, fast and bright
Small cells handle color, form, and sight
Optic nerve bundles every visual tale
Through the chiasm, cortical trail
[Chorus]
Back to front, that's where light must go
Through transparent layers, nice and slow
Front to back, signals make their way
Retina's wired in a backwards display
Back to front, front to back
Neural signals on a zigzag track
[Outro]
From photoreceptors to ganglion's end
Forward through the network signals wend
Axons carry messages to the brain
Back to front, then forward once again
[Final Chorus]
Back to front, that's where light must go
Through transparent layers, nice and slow
Front to back, signals make their way
Retina's wired in a backwards display
Back to front, front to back
Neural signals on a zigzag track
6. The computation: center-surround and what it means
[Verse 1]
Deep in your retina, cells are computing
A difference of Gaussians, signals commuting
Center-surround fields, they're checking the scene
Bright spot on darkness makes neurons convene
ON-center fires when contrast appears
But uniform light just disappears
[Chorus]
DoG computation, center minus surround
Band-pass spatial filter, edges get found
Whitening the spectrum, flattening the curve
Matching optic bandwidth to what we observe
Center-surround magic, contrast is king
Discarding the background, edges take wing
[Verse 2]
Natural images follow one over f-squared
Power spectrum falling, statistics are shared
But center-surround flips that distribution
Decorrelates pixels through smart contribution
Atick and Redlich proved the connection
Information theory meets retinal perfection
[Chorus]
DoG computation, center minus surround
Band-pass spatial filter, edges get found
Whitening the spectrum, flattening the curve
Matching optic bandwidth to what we observe
Center-surround magic, contrast is king
Discarding the background, edges take wing
[Verse 3]
Engineering wisdom hidden in biology
Efficient coding solved through retinal ecology
Channel capacity matched to world statistics
Center-surround proves math isn't just academics
Luminance discarded, local contrast amplified
Nature's compression algorithm, perfectly applied
[Bridge]
Why waste precious axons on redundant data
When you can preprocess, make signals much cleaner
Retinal ganglion cells learned optimization
Through millions of years of neural refinement
Suppressed by uniform, excited by borders
Following information-theoretic orders
[Verse 4]
Two Gaussian kernels, one narrow, one wide
Subtracting their outputs, redundancy dies
Mexican hat filters in visual cortex too
The same DoG principle carries right through
From photoreceptors to higher brain regions
Center-surround math spans neural legions
[Chorus]
DoG computation, center minus surround
Band-pass spatial filter, edges get found
Whitening the spectrum, flattening the curve
Matching optic bandwidth to what we observe
Center-surround magic, contrast is king
Discarding the background, edges take wing
[Outro]
From Gaussian differences to redundancy reduction
Center-surround shows evolution's contribution
Information theory meets retinal design
Where biology and mathematics align
7. What is on the wire leaving the eye
[Verse 1]
When photons hit your retina's stage
The eye don't send what cameras gauge
No pixels marching line by line
Instead it runs a different design
Twenty channels spike and fire
Each one a specialized wire
Ganglion cells sort the scene
Into streams you've never seen
[Chorus]
Twenty channels, twenty views
ON and OFF in different hues
Center-surround begins the tale
Motion, color, never fail
Tiling patterns, complete and wide
Twenty filters side by side
From the eye the story flows
In parallel streams it goes
[Verse 2]
Kuffler back in fifty-three
Found the code we couldn't see
Center bright and surround dark
Or flip the switch, reverse the mark
Each ganglion type spreads its net
Across the retina's silhouette
Complete coverage, overlapping maps
Twenty different neural taps
[Chorus]
Twenty channels, twenty views
ON and OFF in different hues
Center-surround begins the tale
Motion, color, never fail
Tiling patterns, complete and wide
Twenty filters side by side
From the eye the story flows
In parallel streams it goes
[Verse 3]
Color opponents take their stance
Red-green, blue-yellow advance
Transient motion catches change
Direction selectors arrange
Looming signals sound alarm
Differential motion means harm
Specialized channels stand on guard
While others make the basics hard
[Verse 4]
Midget cells for fine detail
Parasol cells that never fail
Y-cells respond to moving edge
X-cells keep their solemn pledge
To linear summation's rule
While Y-cells break it like a fool
Sustained response or transient burst
Each channel knows which comes first
[Bridge]
In the fovea packed so tight
Sampling dense for sharpest sight
Out toward the edge it thins
That's where magnification begins
Not the cortex making space
Retina sets the sampling pace
Best-understood circuit in the brain
Each connection mapped in molecular chain
[Chorus]
Twenty channels, twenty views
ON and OFF in different hues
Center-surround begins the tale
Motion, color, never fail
Tiling patterns, complete and wide
Twenty filters side by side
From the eye the story flows
In parallel streams it goes
[Outro]
Efficient coding theory shows
Why evolution chose
These twenty streams of spiking code
Down the optic nerve they rode
No pixels on this neural wire
Just filtered scenes that never tire
[Final Chorus]
Twenty channels, twenty views
ON and OFF in different hues
Center-surround begins the tale
Motion, color, never fail
From the eye the story flows
In parallel streams it goes
8. The decussation rule
[Verse 1]
Light bounces off the world into your waiting eyes
Through the pupil, hits the lens, gets flipped and twisted twice
What was left becomes the right, what was right becomes the left
The retina receives this topsy-turvy theft
Nasal side and temporal side, each with different fates
One will cross, one will stay, as the visual pathway waits
[Chorus]
At the chiasm crossing, where the optic highways meet
Nasal fibers jump the line, temporal fibers keep their beat
Left brain sees the right side world, right brain takes the left
Contralateral visual fields, perfectly bereft
From both eyes to one hemisphere, the crossing rule's complete
Binocular combination makes the neural circuit sweet
[Verse 2]
Picture standing in a field, mountains on your right
Your left eye's temporal side captures that majestic sight
But your right eye's nasal half sees those same peaks so tall
Both send signals to left brain, answering the call
The lens has done its flipping trick, nasal sees what's far away
Temporal captures close-up views in this neural ballet
[Chorus]
At the chiasm crossing, where the optic highways meet
Nasal fibers jump the line, temporal fibers keep their beat
Left brain sees the right side world, right brain takes the left
Contralateral visual fields, perfectly bereft
From both eyes to one hemisphere, the crossing rule's complete
Binocular combination makes the neural circuit sweet
[Bridge]
Stereopsis needs this plan, depth perception's clever scheme
Corresponding retinal points merge into one stream
Same hemisphere processing, same cortical column too
Two eyes become a single view, dimensional and true
The crossing rule delivers what binocular sight demands
United visual processing from evolution's hands
[Verse 3]
Remember this decussation, half cross and half remain
Nasal axons switch their sides, temporal stay in lane
Each hemisphere receives the whole contralateral scene
From both your eyes combined as one, sharp and crystal clean
The chiasm sorts the traffic with this ancient neural code
Splitting visual information down the proper road
[Chorus]
At the chiasm crossing, where the optic highways meet
Nasal fibers jump the line, temporal fibers keep their beat
Left brain sees the right side world, right brain takes the left
Contralateral visual fields, perfectly bereft
From both eyes to one hemisphere, the crossing rule's complete
Binocular combination makes the neural circuit sweet
[Outro]
So when you see the world around, remember how it flows
Crossed and sorted, mixed and matched, as visual data goes
The decussation rule ensures your hemispheres align
Processing the world you see in patterns so divine
9. Why the routing matters: lesion localization
[Verse 1]
When the brain maps vision, there's a pathway to trace
From retina to cortex, each fiber finds its place
But when damage strikes along this neural highway
The pattern of your blindness tells us where things went astray
One optic nerve gets severed, that eye goes completely dark
While the other keeps on seeing, like a candle in the park
[Chorus]
Cut before the crossing, lose one eye alone
Cut right at the junction, outer fields are gone
Cut behind the meeting, same side disappears
The pattern tells the story that the doctor's trained to hear
Routing matters deeply when the visual system breaks
Every lesion's signature shows the path that vision takes
[Verse 2]
At the chiasm crossing, nasal fibers interweave
Like dancers switching partners in a neurologic weave
When pituitary tumors press from underneath this spot
They crush the crossing fibers but the straight-through ones they're not
Bitemporal hemianopia means both outer fields have fled
While the center and the inside keep the visual pathway fed
[Chorus]
Cut before the crossing, lose one eye alone
Cut right at the junction, outer fields are gone
Cut behind the meeting, same side disappears
The pattern tells the story that the doctor's trained to hear
Routing matters deeply when the visual system breaks
Every lesion's signature shows the path that vision takes
[Verse 3]
Past the chiasm junction, in the tract and LGN
When damage strikes these regions, there's a pattern once again
Homonymous hemianopia takes the same side from both eyes
Contralateral blindness where the damaged pathway lies
But retinotopy's preservation keeps the map intact and clear
From the eyeball to cortex, every point is stationed here
[Bridge]
In the radiations spreading to the primary visual zone
The fovea gets protection from its blood supply alone
Macular sparing happens when the center stays alive
While peripheral vision fades, the detail zone will thrive
This clinical neuroanatomy tells us something profound
The orderly arrangement can be traced and can be found
[Chorus]
Cut before the crossing, lose one eye alone
Cut right at the junction, outer fields are gone
Cut behind the meeting, same side disappears
The pattern tells the story that the doctor's trained to hear
Routing matters deeply when the visual system breaks
Every lesion's signature shows the path that vision takes
[Outro]
From optic nerve to cortex, through chiasm, tract and more
Each lesion writes its story on the visual field's door
The routing tells us everything about where damage lies
In the wondrous neural highway that connects our brain to eyes
10. Where the axons actually go
[Verse 1]
From the retina's ganglion cells, a million axons take their flight
Through the optic tract they journey, splitting left and right
Not one destination waiting, but five pathways to explore
Each with its own sacred mission, each unlocking different doors
[Chorus]
LGN gets ninety percent, the thalamic relay station
Superior colliculus moves your eyes with rapid orchestration
Pretectum shrinks your pupils when the brightness starts to grow
Suprachiasmatic keeps time, while accessory systems flow
The eye speaks many languages, to places high and low
[Verse 2]
The lateral geniculate nucleus, that's where most fibers land
Ninety percent march onward to the cortex so grand
Building conscious visual pictures in areas V1 and more
But that's just the largest pipeline, not the only neural door
[Chorus]
LGN gets ninety percent, the thalamic relay station
Superior colliculus moves your eyes with rapid orchestration
Pretectum shrinks your pupils when the brightness starts to grow
Suprachiasmatic keeps time, while accessory systems flow
The eye speaks many languages, to places high and low
[Verse 3]
Superior colliculus in midbrain territory dwells
Directing eye movements swiftly when danger story tells
Threat response and orienting, saccades both fast and clean
While pretectal nuclei manage the pupillary scene
[Bridge]
Suprachiasmatic nucleus reads the rhythm of the day
Intrinsically photosensitive cells light up the circadian way
Accessory optic pathways with pulvinar's helping hand
Keep your gaze rock steady as you survey the land
[Verse 4]
Parallel processing channels running simultaneously
Not serial computation, but concurrent symphony
Blindsight shows us clearly when cortical paths are broke
Other routes keep functioning, though consciousness is cloaked
[Verse 5]
From chiasm's crossroads emerging, information flows divergent
Each target serves a purpose both ancient and emergent
Evolution built these highways through millions of years
Five streams of visual data, each pathway crystal clear
[Final Chorus]
LGN gets ninety percent, the thalamic relay station
Superior colliculus moves your eyes with rapid orchestration
Pretectum shrinks your pupils when the brightness starts to grow
Suprachiasmatic keeps time, while accessory systems flow
Five destinations total, that's how visual pathways go
[Outro Chorus]
LGN gets ninety percent, the thalamic relay station
Superior colliculus moves your eyes with rapid orchestration
Pretectum shrinks your pupils when the brightness starts to grow
Suprachiasmatic keeps time, while accessory systems flow
The eye speaks many languages, to places high and low
11. The lateral geniculate nucleus (LGN)
[Verse 1]
Deep in the thalamus, a gateway stands
Six layers thick with neural strands
The LGN sorts what your eyes have caught
Before it reaches the cortex thought
Center-surround fields still hold their ground
No radical transforms to be found
Just refining channels, keeping them clear
As visual signals venture near
[Chorus]
One-two magno, three-four-five-six parvo
Konio hides between the layers we know
Contra-ipsi-contra, eyes alternate their flow
One-four-six from the other side go
Two-three-five from your same side show
LGN's the relay, organizing the show
[Verse 2]
Layers one and two house the giants inside
Magnocellular cells with nowhere to hide
Parasol ganglions feed them well
Fast transient signals they love to tell
Achromatic visions, high contrast scenes
Motion and coarse forms, they're processing machines
Big cells for big picture, movement and more
While smaller layers wait for their chore
[Chorus]
One-two magno, three-four-five-six parvo
Konio hides between the layers we know
Contra-ipsi-contra, eyes alternate their flow
One-four-six from the other side go
Two-three-five from your same side show
LGN's the relay, organizing the show
[Verse 3]
Three through six belong to parvocellular crew
Midget cells whisper what they once knew
Sustained responses, red-green debate
Fine spatial details they calculate
Color opponents locked in their dance
Giving small features their fighting chance
Four layers strong for precision and hue
While mysteries lurk in sublayers too
[Verse 4]
From retinal maps to cortical zones
The LGN preserves what topology owns
Retinotopic order stays intact
Visual fields precisely mapped and tracked
Feedback descending from cortex above
Fine-tunes the signals with push and shove
Not just a relay, but active gate
Controlling which messages reach their fate
[Bridge]
Koniocellular secrets tucked away
Blue-yellow channels have their say
Scientists mapping what else they do
In thin sublayers, functions still new
Each eye alone, no fusion yet
That combination cortex will beget
[Chorus]
One-two magno, three-four-five-six parvo
Konio hides between the layers we know
Contra-ipsi-contra, eyes alternate their flow
One-four-six from the other side go
Two-three-five from your same side show
LGN's the relay, organizing the show
[Outro]
Gateway keeper, thalamic throne
Parallel channels, each stands alone
Till cortex calls and the merger begins
That's where the real integration wins
12. The puzzle of the LGN: why is it there at all?
[Verse 1]
Deep in your brain there's a curious station
Between the eye and the visual cortex destination
The LGN sits like a puzzle unsolved
If it barely changes signals, why's it evolved?
Scientists wondered for decades in vain
What's this waypoint doing in the visual chain?
[Chorus]
Five to ten percent from the retina flows
Ninety percent from somewhere else goes
Feedback and gating, arousal control
The LGN's got a different role
Not just a relay, it's a traffic cop smart
Deciding which signals get the green light to start
[Verse 2]
V1 sends massive projections right back
Cortical feedback on the visual track
Thalamic reticular throws inhibition around
While brainstem modulators change the whole sound
Cholinergic and noradrenergic too
Attention and arousal shaping what gets through
[Chorus]
Five to ten percent from the retina flows
Ninety percent from somewhere else goes
Feedback and gating, arousal control
The LGN's got a different role
Not just a relay, it's a traffic cop smart
Deciding which signals get the green light to start
[Bridge]
When you're sleeping it bursts and blocks the stream
No faithful relay in the land of dreams
But when you're alert and paying attention close
Cortical feedback sharpens what it chose
First place in the stack where top-down takes hold
The retina's data, but the brain's in control
[Verse 3]
Gain-controlled, state-dependent gate
Your arousal level seals the visual fate
The eye provides the raw information feed
But higher brain regions decide what you need
This ancient structure's got a modern surprise
It's not what you see, but what your brain buys
[Verse 4]
Context and expectation flow from above
Predictive signals that push and shove
The LGN compares what's expected to arrive
With retinal input to keep perception alive
Error signals and corrections in the mix
Where prediction meets reality, that's where it clicks
[Chorus]
Five to ten percent from the retina flows
Ninety percent from somewhere else goes
Feedback and gating, arousal control
The LGN's got a different role
Not just a relay, it's a traffic cop smart
Deciding which signals get the green light to start
[Outro]
So the puzzle's solved with wiring statistics
The LGN's role is characteristic
Bottom-up data meets top-down command
At this crucial junction where signals are planned
[Final Chorus]
Five to ten percent from the retina flows
Ninety percent from somewhere else goes
The mystery's solved, now we understand
The LGN's the brain's commanding hand
13. Superior colliculus (SC)
[Verse 1]
Deep in the midbrain where the ancient circuits dwell
Lives a structure with a story evolutionists tell
The Superior Colliculus, layered and wise
In fish it's called tectum, the master of eyes
Two levels of function stacked neat and clean
The most primitive vision machine you've ever seen
[Chorus]
SC maps the world in layers, superficial and deep
Retinotopic on top where the visual signals keep
Motor maps below drive the saccades that leap
Fast reflexes guard you even when you sleep
Superior Colliculus, ancient and true
Spinning your eyeballs to capture the view
[Verse 2]
The shallow layers paint a picture precise
Every spot on your retina gets mapped out nice
Point to point like a blueprint of sight
But venture deeper and you'll find something bright
Motor neurons fire when targets appear
Commanding quick saccades to bring objects near
[Chorus]
SC maps the world in layers, superficial and deep
Retinotopic on top where the visual signals keep
Motor maps below drive the saccades that leap
Fast reflexes guard you even when you sleep
Superior Colliculus, ancient and true
Spinning your eyeballs to capture the view
[Bridge]
Sound and touch join the party down low
Multisensory fusion puts on quite a show
Head coordinates, eye coordinates align
When danger approaches, the response is divine
Blindsight patients can point to what they can't see
Their cortex is broken but SC runs free
[Verse 3]
Orienting responses happen lightning fast
Before consciousness knows, the moment has passed
Salient stimuli trigger the dance
Of microsaccades in a reflexive trance
From optic tectum to colliculus grand
Evolution's gift helps survival withstand
[Verse 4]
Inhibition and excitation compete in the race
Each neuron voting for where eyes should place
Winner takes all in this neural debate
While losers are silenced to seal their fate
The built-in GPS of your visual brain
Keeps tracking on target through sun and through rain
[Chorus]
SC maps the world in layers, superficial and deep
Retinotopic on top where the visual signals keep
Motor maps below drive the saccades that leap
Fast reflexes guard you even when you sleep
Superior Colliculus, ancient and true
Spinning your eyeballs to capture the view
[Outro]
When sudden motion catches your gaze
Thank the colliculus for those ancient ways
Layered maps working in perfect time
The midbrain's rhythm, the visual rhyme
14. Pulvinar
[Verse 1]
Deep inside your thalamus lies a giant of the brain
The pulvinar nucleus, the largest in the chain
While other relays pass things through like messengers in line
This one's got a different job, more subtle and refined
[Chorus]
Pulvinar synchronizer, cortico-cortical relay
Timing all the chatter that your cortex speaks each day
Visual attention keeper, filtering what's not right
When the pulvinar's damaged, you lose half your sight
[Verse 2]
Two cortical regions want to share their latest news
Instead of talking face to face, the pulvinar they use
It choreographs the timing, makes the signals synchronized
Regulates the salience of what gets recognized
[Chorus]
Pulvinar synchronizer, cortico-cortical relay
Timing all the chatter that your cortex speaks each day
Visual attention keeper, filtering what's not right
When the pulvinar's damaged, you lose half your sight
[Verse 3]
Picture walking through a crowd, distractors everywhere
Your pulvinar's the bouncer that decides what gets through where
It amplifies what matters while it dims the background noise
Without it working properly, you'd have no visual choice
[Verse 4]
From parietal to temporal, from frontal back to V1
The pulvinar connects them all until the job is done
It modulates the gamma waves that bind perception tight
Creating conscious awareness from scattered points of light
[Bridge]
Scientists debate its role, the function's not quite clear
But evidence keeps mounting that attention lives right here
Neglect syndrome patients show us what we stand to lose
When pulvinar connections get severed, bruised, or confused
[Chorus]
Pulvinar synchronizer, cortico-cortical relay
Timing all the chatter that your cortex speaks each day
Visual attention keeper, filtering what's not right
When the pulvinar's damaged, you lose half your sight
[Verse 5]
So remember this large nucleus when studying the brain
It's not just passing messages like links in a data chain
It's orchestrating symphonies of cortical debate
Making sure your visual world stays sharp and up to date
[Chorus]
Pulvinar synchronizer, cortico-cortical relay
Timing all the chatter that your cortex speaks each day
Visual attention keeper, filtering what's not right
When the pulvinar's damaged, you lose half your sight
[Outro]
In the thalamus it's waiting, doing jobs we're learning still
The mighty pulvinar nucleus, master of attention's will
15. Pretectum and SCN
[Verse 1]
Deep in your midbrain, two stations await
Processing light signals that seal your fate
The pretectum watches for brightness to shift
While SCN guards time's circadian drift
When photons hit retina, messages race
Through optic nerve pathways to each special place
[Chorus]
Pretectum makes pupils dance in sync
SCN sets your body's clock in sync
Light reflex and rhythm, they both do their thing
Two guardians of sight that the midbrain will bring
Test the pupil response when the cortex won't think
Pretectum and SCN working in sync
[Verse 2]
Shine light in one eye, watch both pupils shrink
That's consensual constriction, faster than you blink
The pretectum's calling shots without asking your brain
Clinical workhorse that doctors maintain
Testing retina-to-midbrain when higher centers fail
This reflex arc tells a neurological tale
[Chorus]
Pretectum makes pupils dance in sync
SCN sets your body's clock in sync
Light reflex and rhythm, they both do their thing
Two guardians of sight that the midbrain will bring
Test the pupil response when the cortex won't think
Pretectum and SCN working in sync
[Verse 3]
Special ganglion cells, intrinsically bright
Send ipRGC signals about day and night
To suprachiasmatic nucleus they relay
Information that keeps circadian sway
Not image vision, but essential for life
Entraining your rhythms through darkness and light
[Verse 4]
When trauma strikes hard and consciousness fades
These ancient pathways still make the grades
Pupillary light reflex remains intact
A brainstem function that stays on track
While cortical vision may be lost to injury
Pretectum and SCN maintain their synergy
[Bridge]
When doctors shine flashlights in emergency rooms
They're testing pretectum to banish their gloom
While SCN whispers the time of day
Through melanopsin's ancient display
Two functions of eyesight beyond what we see
Reflex and rhythm working perfectly
[Chorus]
Pretectum makes pupils dance in sync
SCN sets your body's clock in sync
Light reflex and rhythm, they both do their thing
Two guardians of sight that the midbrain will bring
Test the pupil response when the cortex won't think
Pretectum and SCN working in sync
[Chorus]
Pretectum makes pupils dance in sync
SCN sets your body's clock in sync
Light reflex and rhythm, they both do their thing
Two guardians of sight that the midbrain will bring
Test the pupil response when the cortex won't think
Pretectum and SCN working in sync
[Outro]
From retina's signals to midbrain's embrace
These stations process light with elegant grace
16. Laminar structure and the canonical microcircuit
[Verse 1]
Six layers stack like floors in V1's tower
Layer four receives the visual power
From LGN the signals come streaming
Magno hits four C alpha, parvo four C beta gleaming
Center-surround circles, monocular and plain
One eye at a time, no tilt to explain
[Chorus]
Four takes input, two-three transform
Five-six output in cortical form
Monocular becomes binocular sight
Circles turn to edges burning bright
This is the circuit, canonical and true
V1's magic making vision new
[Verse 2]
Layer four projects its data upward bound
To superficial two and three where orientation's found
Binocular fusion happens in this zone
Both eyes together, no longer alone
Feedforward highways carry signals ahead
To higher cortex where complex forms are read
[Chorus]
Four takes input, two-three transform
Five-six output in cortical form
Monocular becomes binocular sight
Circles turn to edges burning bright
This is the circuit, canonical and true
V1's magic making vision new
[Bridge]
Deep layers five and six control the flow
Five to superior colliculus below
Six sends feedback to LGN's gate
Modulating four's receptive state
The circuit loops, a neural conversation
Each layer plays its role in visual sensation
[Verse 3]
Unoriented blobs become tilted lines
Center-surround transforms to border designs
Single eye input splits to double vision
V1's computation makes the key decision
From simple circles to complex detection
This laminar flow builds our perception
[Verse 4]
Inhibition sculpts the neural dance
GABAergic cells enhance the signal's chance
Chandelier and basket cells refine
The orientation columns, sharp and fine
Horizontal connections spread across
Linking similar angles, preventing loss
[Chorus]
Four takes input, two-three transform
Five-six output in cortical form
Monocular becomes binocular sight
Circles turn to edges burning bright
This is the circuit, canonical and true
V1's magic making vision new
[Chorus]
Four takes input, two-three transform
Five-six output in cortical form
Monocular becomes binocular sight
Circles turn to edges burning bright
This is the circuit, canonical and true
V1's magic making vision new
[Outro]
Six layers working, each with its part
The canonical circuit, visual art
From LGN to cortex above
That's how the brain learns what to love
17. The two cell types Hubel and Wiesel found
[Verse 1]
In the visual cortex, two pioneers found
Different types of neurons, vision-processing bound
Hubel and Wiesel mapped the neural terrain
Simple cells first, let me now explain
They love edges and bars in one direction
At precise positions, no room for correction
Gabor filters model how they operate
Gaussian envelope with sine waves ornate
[Chorus]
Simple cells detect, complex cells collect
Filter then pool, that's the architect
ON and OFF regions, side by side aligned
Position matters to the simple mind
But complex cells don't care where you place it
Phase invariant, they'll never waste it
Filter then pool, the motif so true
CNNs copied what biology knew
[Verse 2]
Simple cells have regions, ON and OFF domains
Laid out side by side like railroad trains
Move that stimulus just a tiny bit
Phase-sensitive response will quickly quit
LGN center-surround inputs combine
Creating orientation preference so fine
Like aligned spotlights in formation
Building edge detection across the nation
[Chorus]
Simple cells detect, complex cells collect
Filter then pool, that's the architect
ON and OFF regions, side by side aligned
Position matters to the simple mind
But complex cells don't care where you place it
Phase invariant, they'll never waste it
Filter then pool, the motif so true
CNNs copied what biology knew
[Verse 3]
Complex cells take a different approach
Same orientation, broader scope coach
Position invariant across their field
Bar can wander, they'll never yield
Energy model explains their trick
Square and sum, the math's not thick
Quadrature simple cells pooled together
Building invariance, smart and clever
[Verse 4]
From cat experiments to monkey sight
Recording electrodes revealed the light
Orientation columns, pinwheel arrays
Hypercolumns mapping visual space for days
Nobel Prize winners showed us the way
How cortical circuits process and display
The building blocks of what we see
Visual hierarchy's first degree
[Bridge]
From the Neocognitron to modern day
Convolutional networks learned the way
Max pooling, average pooling too
All inspired by what these neurons do
Biological blueprint, engineered design
V1's architecture, perfectly refined
[Chorus]
Simple cells detect, complex cells collect
Filter then pool, that's the architect
ON and OFF regions, side by side aligned
Position matters to the simple mind
But complex cells don't care where you place it
Phase invariant, they'll never waste it
Filter then pool, the motif so true
CNNs copied what biology knew
[Outro]
Two cell types in the cortical dance
Hubel and Wiesel gave us the chance
To understand how vision starts to grow
From simple features to complex show
18. Other V1 feature dimensions
[Verse 1]
Beyond the bars and edges that we know so well
V1's got secrets, more stories to tell
Spatial frequency sorts the coarse from fine
Rough bark from silk, each texture's divine
Some neurons love the details, crisp and tight
Others prefer the blur, the broader sight
[Chorus]
Five dimensions dancing in your visual cortex
Spatial frequency, motion, depth connects
Binocular disparity brings the world alive
Color blobs and end-stopping help us thrive
V1's not just orientation, there's so much more
Five feature filters at the neural core
[Verse 2]
Motion detectors catch the leftward flow
Magno-driven cells know which way things go
Direction selective, they fire with precision
Up, down, diagonal - each has its mission
While stationary patterns leave them cold
Moving targets make their stories unfold
[Chorus]
Five dimensions dancing in your visual cortex
Spatial frequency, motion, depth connects
Binocular disparity brings the world alive
Color blobs and end-stopping help us thrive
V1's not just orientation, there's so much more
Five feature filters at the neural core
[Verse 3]
Two eyes diverge but V1 makes them one
Binocular disparity's where depth's begun
Left eye, right eye, slightly different views
This tiny shift gives stereoscopic clues
First cortical stage where the eyes combine
Turning flat pictures into depth divine
[Verse 4]
Hypercolumns organize this neural maze
Each contains the maps that guide our gaze
Orientation columns spinning all around
While ocular preference keeps eyes unbound
Six layers deep in cortical design
Simple, complex cells in ordered lines
[Bridge]
Cytochrome oxidase blobs light up the scene
Processing color in patches of protein
Red, blue, and yellow find their neural home
While end-stopped cells won't let long bars roam
Fire for short lines but shut down when extended
Curvature detection, context blended
[Chorus]
Five dimensions dancing in your visual cortex
Spatial frequency, motion, depth connects
Binocular disparity brings the world alive
Color blobs and end-stopping help us thrive
V1's not just orientation, there's so much more
Five feature filters at the neural core
[Outro]
From texture maps to motion's flow
From stereo depth to color's glow
End-stopping curves complete the show
V1's dimensions help us know
[Final Chorus]
Five dimensions dancing in your visual cortex
Spatial frequency, motion, depth connects
Binocular disparity brings the world alive
Color blobs and end-stopping help us thrive
V1's not just orientation, there's so much more
Five feature filters at the neural core
19. The functional architecture: columns, hypercolumns, blobs, pinwheels
[Verse 1]
Deep in V1 there's a masterpiece design
Columns standing tall in cortical lines
Cells that share the same direction they adore
Stacked up vertical through each neural floor
Orientation preference spins around
Like a compass wheel upon the visual ground
[Chorus]
Hypercolumns mapping every point of sight
Blobs and pinwheels dancing in the cortical light
Left eye, right eye, stripes of dominance flow
Cytochrome burning where the colors grow
One to two millimeters of processing power
V1's modular magic every waking hour
[Verse 2]
Hubel and Wiesel showed us zebra bands
Autoradiographs revealing nature's plans
Ocular dominance in alternating streams
Left and right eye channels weaving visual dreams
Cover up one eye during critical time
Watch the plastic changes rearrange the rhyme
[Chorus]
Hypercolumns mapping every point of sight
Blobs and pinwheels dancing in the cortical light
Left eye, right eye, stripes of dominance flow
Cytochrome burning where the colors grow
One to two millimeters of processing power
V1's modular magic every waking hour
[Bridge]
Pinwheel singularities mark the spots where
Orientation maps spiral everywhere
Optical imaging shows the living brain
Two-photon microscopes reveal the hidden grain
Blobs handle hue while interblobs process form
Color versus edges in the visual storm
[Verse 3]
Cytochrome oxidase marks the metabolic pegs
Enzyme-rich regions like barrel-shaped legs
Color-loving neurons cluster in these zones
Orientation-blind, they work in different tones
Form and motion live between the blob arrays
Processing boundaries in complementary ways
[Verse 4]
Development unfolds through activity waves
Spontaneous firing that the cortex craves
Layer four receives the thalamic drive
While layers five and six keep feedback alive
Inhibition sculpts the receptive field shape
Gamma oscillations help the columns escape
[Chorus]
Hypercolumns mapping every point of sight
Blobs and pinwheels dancing in the cortical light
Left eye, right eye, stripes of dominance flow
Cytochrome burning where the colors grow
One to two millimeters of processing power
V1's modular magic every waking hour
[Outro]
Retinotopic tiles across the cortical sheet
Making vision possible, the cycle complete
Hypercolumns repeat their sacred song
Processing images all day long
[Final Chorus]
Hypercolumns mapping every point of sight
Blobs and pinwheels dancing in the cortical light
Left eye, right eye, stripes of dominance flow
V1's modular magic every waking hour
20. Retinotopy and cortical magnification
[Verse 1]
In the back of your brain where the vision starts flowing
V1's got a map that keeps perfectly showing
Every point from your eye finds its place on the wall
Neighbors stay as neighbors through cortical sprawl
But this map's got a secret, a twist in its plan
The fovea's blown up like a zoom-in span
[Chorus]
Retinotopic mapping, point to point it goes
But cortical magnification, that's how the story flows
Half of V1 devoted to ten degrees of sight
Foveal cones packed so densely, they steal the spotlight bright
Map it out, blow it up, central vision wins the prize
That's why your periphery can't match your focused eyes
[Verse 2]
Hubel and Wiesel with their electrodes fine
Traced the single units, mapped them line by line
Optical imaging made the columns glow
Two-photon calcium let the patterns show
fMRI in humans matched the monkey's tale
Magnification factors, precise without fail
[Chorus]
Retinotopic mapping, point to point it goes
But cortical magnification, that's how the story flows
Half of V1 devoted to ten degrees of sight
Foveal cones packed so densely, they steal the spotlight bright
Map it out, blow it up, central vision wins the prize
That's why your periphery can't match your focused eyes
[Bridge]
When V1 gets damaged, vision disappears
Scotomas prove the link through clinical years
Midget cells oversample the central zone
Filter-then-pool models, quantitatively shown
The strongest evidence base the field has ever known
From single cells to human brains, the patterns clearly shown
[Verse 3]
Acuity collapses as you move away
From that precious center where the details stay
Cortical real estate commands a hefty price
Central ten degrees get the neural slice
Pinwheels spinning, columns in their rows
Following the blueprint that biology chose
[Verse 4]
Evolution shaped this neural architecture
For reading faces, finding food for sure
Sharp detail in the center, motion on the sides
Cortical commitment where each function hides
M squared per degree, the formula's clean
Mathematical beauty in the vision machine
[Final Chorus]
Retinotopic mapping, point to point it goes
But cortical magnification, that's how the story flows
Half of V1 devoted to ten degrees of sight
Foveal cones packed so densely, they steal the spotlight bright
Map it out, blow it up, central vision wins the prize
That's the cortical echo behind your seeing eyes
21. The original formulation (Ungerleider and Mishkin, 1982): "what" vs "where"
[Verse 1]
Back in eighty-two, two scientists had a hunch
Studied monkey brains with surgical precision's punch
Ungerleider and Mishkin made lesions clean and neat
Found two streams of vision flowing from cortical beat
V1 starts the journey, splits the visual load
Down the ventral highway or the dorsal road
[Chorus]
What pathway travels down, where pathway climbs up high
Ventral stream says "what is it," dorsal asks "where am I"
From V1 to V2, then the streams divide
What goes to temporal, where to parietal side
Two questions that your brain keeps asking every day
What am I seeing and where does it lay
[Verse 2]
Ventral stream cascades through V2 then V4's gate
Inferotemporal cortex seals the object's fate
Cut this pathway and the monkey's vision breaks apart
Sees the banana but can't name what fills its heart
Recognition vanishes though the eyes still work fine
The what pathway's damage cuts the meaning line
[Chorus]
What pathway travels down, where pathway climbs up high
Ventral stream says "what is it," dorsal asks "where am I"
From V1 to V2, then the streams divide
What goes to temporal, where to parietal side
Two questions that your brain keeps asking every day
What am I seeing and where does it lay
[Verse 3]
Dorsal stream climbs upward through V2 and V3's maze
MT processes motion in the visual haze
Posterior parietal cortex maps the spatial scene
Damage here destroys location's crystalline sheen
Monkey knows it's food but can't reach out to grab
The where pathway's broken, spatial sense gone bad
[Verse 4]
Human patients show the same divided sight
Stroke victims lose their what or where overnight
Visual agnosia when ventral stream is hurt
Can't name a spoon but know where objects assert
Optic ataxia from dorsal damage shows
Reaching for a cup but missing where it goes
[Bridge]
Temporal lobe remembers faces, objects, forms
Parietal lobe navigates through spatial storms
One stream for identity, one for position
Ancient pathways serving visual cognition
Lesion studies revealed this dual design
Two streams of consciousness, yours and mine
[Chorus]
What pathway travels down, where pathway climbs up high
Ventral stream says "what is it," dorsal asks "where am I"
From V1 to V2, then the streams divide
What goes to temporal, where to parietal side
Two questions that your brain keeps asking every day
What am I seeing and where does it lay
[Outro]
So remember the split when you open your eyes
What versus where, the brain's great disguise
Processing vision through parallel lanes
Ungerleider-Mishkin lives on in our brains
22. The influential reframing (Goodale and Milner, 1992): "what" vs "how"
[Verse 1]
Back in ninety-two, Goodale and Milner saw
Two streams of vision with a different call
Not just "where" and "what" like we thought before
But "how" and "what" - they opened up the door
Dorsal stream computes for action's need
Ventral builds the world that we perceive
[Chorus]
What versus how, the brain divides the load
Vision for perception down the ventral road
Vision for action takes the dorsal way
Two separate systems working night and day
What versus how, remember this refrain
Double dissociation proves they're not the same
[Verse 2]
Meet patient D.F., carbon monoxide's curse
Damaged ventral stream, but here's what's perverse
She cannot tell you if a slot runs straight
Or recognize the size upon her plate
But watch her hand move with perfect grace
Threading cards through slots she cannot trace
[Chorus]
What versus how, the brain divides the load
Vision for perception down the ventral road
Vision for action takes the dorsal way
Two separate systems working night and day
What versus how, remember this refrain
Double dissociation proves they're not the same
[Verse 3]
Optic ataxia tells the flip side tale
Parietal damage makes their reaching fail
They'll name each object, know it crystal clear
But grasp for coffee cups that aren't quite there
Perception's perfect, action goes astray
When dorsal pathways lose their guiding way
[Verse 4]
Temporal cortex for the what we know
Parietal lobe for how our actions flow
Online control versus conscious sight
One works in darkness, one needs the light
Evolution's answer to survive and thrive
Keep these systems functionally alive
[Bridge]
Egocentric coordinates for grasping motion
Allocentric maps for recognition's potion
Real-time guidance versus memory's store
Two parallel worlds behind perception's door
Cleanest evidence that neuroscience has found
Functional separation, safe and sound
[Chorus]
What versus how, the brain divides the load
Vision for perception down the ventral road
Vision for action takes the dorsal way
Two separate systems working night and day
What versus how, remember this refrain
Double dissociation proves they're not the same
[Outro]
So when you reach out for your morning brew
Remember streams of vision, split in two
Goodale and Milner showed us something true
The brain's not one, but systems numbering two
23. The caveats you must keep
[Verse 1]
Two highways in your skull they say, ventral what and dorsal where
But hold your horses, slow your roll, the truth needs careful care
These streams ain't rivers running solo, they're chatting all day long
Cross-connected at each junction, information flowing strong
[Chorus]
Keep your caveats close, don't let assumptions lead you wrong
The two-stream map's a starting point, not the final song
Cross-connected, coarse and messy, monkey brains ain't quite the same
Use it as your routing compass, not your neuron-naming game
[Verse 2]
That dichotomy sounds so clean and neat, but dig a little deeper down
Each stream's got specialized neighborhoods, not just one side of town
Some jobs like three-dimensional shapes recruit from both the crews
So don't go boxing every function into tidy what-and-where news
[Chorus]
Keep your caveats close, don't let assumptions lead you wrong
The two-stream map's a starting point, not the final song
Cross-connected, coarse and messy, monkey brains ain't quite the same
Use it as your routing compass, not your neuron-naming game
[Verse 3]
Now monkey studies paved the way, but humans took a different turn
Parietal cortex stretched and shifted, new lessons we had to learn
Homology ain't perfect matching, evolution played some tricks
So primate papers tell part of the tale, but human brains got different licks
[Verse 4]
When lesions hit the parietal zone, the symptoms tell a complex tale
Not just pure spatial navigation that comes grinding to a fail
Visual attention gets all scrambled, reaching movements go askew
But object recognition lingers, so the what-stream's pushing through
[Bridge]
Where and how ain't fighting rivals, they're dancing partners on the floor
Dorsal stream does spatial mapping and action planning, maybe more
Sub-areas split the workload up, each corner plays its part
Don't pin it all on single cells, that's missing the art
[Chorus]
Keep your caveats close, don't let assumptions lead you wrong
The two-stream map's a starting point, not the final song
Cross-connected, coarse and messy, monkey brains ain't quite the same
Use it as your routing compass, not your neuron-naming game
[Outro]
High-level routing, that's the game
Single neurons? Not the same
Keep those caveats burning bright
In your visual processing sight
24. V2 - second visual area
[Verse 1]
After V1 does its primary scan
V2 steps up with a grander plan
The largest area downstream from sight
Organized in stripes of cellular might
Cytochrome oxidase paints the way
Thick and thin and pale display
Magno parvo blob segregation
Continues through this brain location
[Chorus]
V2 sees what isn't there
Illusory contours in the air
Border ownership decides the ground
Figure from background can be found
Stripes of thick carry motion and depth
Thin stripes handle color with adept
Pale stripes work on form and line
V2 makes the abstract shine
[Verse 2]
Kanizsa triangles fool the eye
But V2 neurons still reply
To edges that don't really exist
Context clues they can't resist
The system fills in structure bold
From gaps and hints the story's told
No luminance edge upon the scene
Yet boundaries clear and crisp are seen
[Chorus]
V2 sees what isn't there
Illusory contours in the air
Border ownership decides the ground
Figure from background can be found
Stripes of thick carry motion and depth
Thin stripes handle color with adept
Pale stripes work on form and line
V2 makes the abstract shine
[Bridge]
Beyond the classical field it reaches
Feedback loops and lateral teaches
Which side owns this visual border
Figure ground in perfect order
Angles corners curvature too
Complex shapes begin to brew
Combinations of orientations
First steps toward representations
[Verse 3]
Thick stripes process disparity
Motion flows with clarity
Thin stripes catch the color streams
Red and blue in vivid dreams
Pale stripes trace each orientation
Building blocks of shape creation
Three pathways working in concert
Visual processing expert
[Verse 4]
Surface properties V2 defines
Texture brightness subtle signs
Lightness constancy holds true
Despite the shadows passing through
Spatial frequency tuning tight
Both low and high in perfect sight
Harmonic relationships unfold
Complex patterns to behold
[Chorus]
V2 sees what isn't there
Illusory contours in the air
Border ownership decides the ground
Figure from background can be found
Stripes of thick carry motion and depth
Thin stripes handle color with adept
Pale stripes work on form and line
V2 makes the abstract shine
[Outro]
From V1's simple edge detection
V2 builds the next connection
Subjective boundaries take their place
In the visual processing race
25. V4 - intermediate shape and color
[Verse 1]
Deep in V4, where the magic unfolds
Colors stay constant as the story gets told
When shadows shift and lighting does change
This neural region keeps hues in their range
Not raw cone signals, but perceived display
Zeki found neurons that work this way
Track surface color through illuminant shift
V4's the master of this perceptual gift
[Chorus]
V4's the bridge in the visual stream
Color constancy and curvature's dream
Between simple edges and objects complete
Where attention reshapes what neurons will meet
Intermediate features, the middle ground zone
Where visual processing finds its backbone
[Verse 2]
Curvature coding at angular spots
Pasupathy Connor mapped these neural plots
Contour fragments dance in receptive fields
Concentric spirals and radial wheels
Not Cartesian gratings but patterns that bend
Shape parts assembly from beginning to end
Between V1's edges and IT's whole face
V4 builds features in this crucial space
[Chorus]
V4's the bridge in the visual stream
Color constancy and curvature's dream
Between simple edges and objects complete
Where attention reshapes what neurons will meet
Intermediate features, the middle ground zone
Where visual processing finds its backbone
[Verse 3]
Moran Desimone revealed the key
Two stimuli competing for what cells see
But when attention focuses on just one
The ignored stimulus becomes undone
Receptive fields reshape their boundaries tight
Responding as if there's only one sight
Spatial and feature attention converge
V4's where selective focus will emerge
[Verse 4]
From retinal maps to topographic space
V4 maintains each visual place
Foveal magnification still holds true
While complex computations filter through
Multiple maps for color, form, and more
Parallel channels from V2's door
Integration happens in this neural hub
Where vision's puzzle pieces start to rub
[Bridge]
When V4's damaged, achromatopsia appears
Color perception simply disappears
World turns to grayscale, vibrance is lost
Without V4's work, that's the visual cost
From illuminant correction to curvature code
Attention's spotlight on this neural road
[Chorus]
V4's the bridge in the visual stream
Color constancy and curvature's dream
Between simple edges and objects complete
Where attention reshapes what neurons will meet
Intermediate features, the middle ground zone
Where visual processing finds its backbone
[Outro]
Richer responses, harder to define
Mid-level processing in the visual design
V4 stands guard at the crucial divide
Where simple meets complex, side by side
[Chorus]
V4's the bridge in the visual stream
Color constancy and curvature's dream
Between simple edges and objects complete
Where attention reshapes what neurons will meet
Intermediate features, the middle ground zone
Where visual processing finds its backbone
26. Inferotemporal cortex (IT) - object identity
[Verse 1]
Deep in the ventral stream's final station
Lives the IT cortex, object's destination
Large receptive fields span the visual scene
Trading position data for what objects mean
Posterior to anterior, the pathway flows
From TEO to TE, where recognition grows
[Chorus]
IT solves the puzzle, breaks the code
Invariant neurons on this road
Size and position, lighting too
Can't fool these cells, they see right through
Linear decode in hundred milliseconds flat
Object identity, imagine that
[Verse 2]
Tanaka mapped the columns, features combined
Complex shapes and textures, neurally aligned
Individual neurons fire for specific things
Across transformations, the same response rings
A coffee cup tilted, moved left or right
Same IT neuron keeps it in sight
[Chorus]
IT solves the puzzle, breaks the code
Invariant neurons on this road
Size and position, lighting too
Can't fool these cells, they see right through
Linear decode in hundred milliseconds flat
Object identity, imagine that
[Verse 3]
DiCarlo proved with hundreds of cells recording
Simple classifier, human-like decoding
Population vectors tell the visual tale
First hundred milliseconds, they never fail
FFA for faces, patches interconnected
Tsao and Freiwald's network, view-invariant perfected
[Bridge]
PPA reads the places and spatial layout
EBA sees bodies without any doubt
VWFA recycles for letters and words
Culture reshaping what vision has heard
Face space dimensions, shape and appearance
Low-dimensional coding with neural adherence
[Verse 4]
The grandmother cell debate still echoes loud
Single concept neurons in the MTL crowd
Quiroga found Jennifer Aniston's cell
Photos, names, written words ring her bell
From simple features to complex recognition
IT cortex completes the visual mission
[Verse 5]
Visual plasticity shapes the neural code
Experience sculpts each learning episode
Development builds these recognition skills
Through training, feedback, the pathway fills
Critical periods when connections form
IT architecture becomes the norm
[Chorus]
IT solves the puzzle, breaks the code
Invariant neurons on this road
Size and position, lighting too
Can't fool these cells, they see right through
Linear decode in hundred milliseconds flat
Object identity, imagine that
[Outro]
Ventral stream's apex, the final stage
Where visual objects take center stage
IT cortex wisdom, the brain's design
Object recognition, so refined
[Chorus]
IT solves the puzzle, breaks the code
Invariant neurons on this road
Size and position, lighting too
Can't fool these cells, they see right through
Linear decode in hundred milliseconds flat
Object identity, imagine that
27. V3 and V3A
[Verse 1]
Deep in the cortex where vision unfolds
V3 and V3A process stories untold
Between the early areas and motion streams
They bridge the gap with transitional schemes
Taking static shapes and breathing life in
Where depth and movement analysis begins
[Chorus]
V3A responds to motion and disparity
Dynamic forms dance with such clarity
On-ramp to motion, the pathway flows
From static vision to movement that glows
Three-dimensional depth perception's key
V3A unlocks what the eyes truly see
[Verse 2]
Less crisp than neighbors V1 and V2
But V3's purpose breaks clearly through
Form processing with a temporal twist
Dynamic patterns that can't be missed
While V4 handles color down the line
V3 areas make moving objects shine
[Chorus]
V3A responds to motion and disparity
Dynamic forms dance with such clarity
On-ramp to motion, the pathway flows
From static vision to movement that glows
Three-dimensional depth perception's key
V3A unlocks what the eyes truly see
[Bridge]
Binocular rivalry finds its place
When left and right eyes don't quite match their space
V3A calculates the difference between
Creating depth from the disparity scene
Motion sensitivity lights up the zone
When objects shift through space we've known
[Verse 3]
Transitional areas serve their role
Connecting pieces to make vision whole
From orientation bars to complex forms
Through V3 processing the image transforms
Ready for MT and the motion cascade
Where temporal sequences are displayed
[Verse 4]
Neural circuits fire in perfect time
Processing edges both subtle and prime
Speed and direction become crystal clear
As V3A maps what's moving near
Texture gradients reveal surface flow
Making three dimensions start to show
[Chorus]
V3A responds to motion and disparity
Dynamic forms dance with such clarity
On-ramp to motion, the pathway flows
From static vision to movement that glows
Three-dimensional depth perception's key
V3A unlocks what the eyes truly see
[Outro]
In the hierarchy climbing high
V3 areas help us see and fly
Through dimensional space with perfect grace
Motion and depth find their rightful place
[Final Chorus]
V3A responds to motion and disparity
Dynamic forms dance with such clarity
On-ramp to motion, the pathway flows
From static vision to movement that glows
Three-dimensional depth perception's key
V3A unlocks what the eyes truly see
28. MT / V5 - the motion area
[Verse 1]
Deep in cortex where the magic flows
V5 neurons dance in perfect rows
Direction columns lined up neat and tight
Reading motion left and right
They don't care what color fills the scene
Or if the shape is square or lean
Just speed and direction, that's their game
Magnocellular fast, they stake their claim
[Chorus]
MT solves what V1 cannot see
Aperture problem holds no mystery
Local signals merge to global truth
Motion integration, solid proof
Newsome showed us how perception flows
When microstimulation glows
Bias the column, bend the mind
Neural causation redefined
[Verse 2]
V1 sees edges through a tiny door
Component motion, nothing more
A line that slides across the view
Perpendicular is all they knew
But MT gathers signals wide
Integrates from every side
True direction emerges clear
Global motion appears here
[Chorus]
MT solves what V1 cannot see
Aperture problem holds no mystery
Local signals merge to global truth
Motion integration, solid proof
Newsome showed us how perception flows
When microstimulation glows
Bias the column, bend the mind
Neural causation redefined
[Verse 3]
Random dots with varied coherence
Monkeys trained with perseverance
Record the cells that match the task
Neural correlates unmasked
Trial by trial the data shows
MT firing predicts where attention goes
Population votes decide the choice
Neurons speak with unified voice
[Bridge]
Then came the spark, the golden test
Electrode placed with careful quest
Stimulate the chosen site
Push perception left or right
Correlation proved the link
Causation made the skeptics think
Gold standard born that fateful day
Neural code on full display
[Verse 4]
Columnar maps like V1's design
Orientation's paradigm
But here it's direction that's arranged
Motion preferences exchanged
Neighboring cells prefer the same
Angular vectors in the frame
Architecture tells the tale
How motion signals never fail
[Verse 5]
Lesion studies sealed the deal
Damage MT, motion's not real
Patients lost in moving world
Motion blindness now unfurled
Baseball pitches disappear
Cars materializing near
Function follows form so true
Area MT, tried and true
[Outro]
From single cells to what we see
MT bridges mystery
Best understood beyond V1
Motion processing, second to none
[Final Chorus]
MT solves what V1 cannot see
Aperture problem holds no mystery
Local signals merge to global truth
Motion integration, solid proof
29. MST (medial superior temporal)
[Verse 1]
Deep inside your temporal lobe there's a zone called MST
Takes the signals from MT and builds a bigger mystery
When you're walking down the hallway, driving through the countryside
All that motion flowing past you gets processed and refined
[Chorus]
MST reads the optic flow
Expansion, contraction, rotation show
Where you're heading, how you're moving through space
Motion patterns guide your path and stabilize your gaze
MST, medial superior T
Downstream processing, setting your body free
[Verse 2]
Picture standing in a tunnel with the walls expanding wide
That's the signal MST loves when you're moving straight inside
Spiral patterns, swirling vortex, when your head begins to turn
These are global motion signals that your brain has got to learn
[Chorus]
MST reads the optic flow
Expansion, contraction, rotation show
Where you're heading, how you're moving through space
Motion patterns guide your path and stabilize your gaze
MST, medial superior T
Downstream processing, setting your body free
[Bridge]
Vestibular whispers secrets from your inner ear canal
MST integrates the message, visual-vestibular
Self-motion computation, heading estimation too
Large-field motion detectors working hard for me and you
[Verse 3]
Not just single dots or edges like the cells we've seen before
MST wants bigger pictures, motion spread across the floor
Locomotion navigation, keeping balance when you move
Ancient circuits in your cortex help you find your forward groove
[Verse 4]
When the pilot banks the aircraft or the sailor rides the swell
MST is calculating every motion, working well
Complex scenes with multiple objects moving all around
Your MST sorts signal from the noise without a sound
Focus of expansion tells you where your path will lead
This cortical area gives you everything you need
[Chorus]
MST reads the optic flow
Expansion, contraction, rotation show
Where you're heading, how you're moving through space
Motion patterns guide your path and stabilize your gaze
MST, medial superior T
Downstream processing, setting your body free
[Outro]
From MT to MST, the processing chain unfolds
Optic flow and heading signals, stories that your brain has told
30. Posterior parietal cortex (PPC)
[Verse 1]
Deep in the parietal cortex lies a grand command center
Where vision meets with action, where the dorsal stream will enter
From retina to fingertips, it transforms what the eyes can see
Eye-centered coordinates become hand and body geography
[Chorus]
PPC's the bridge between, where you look and what you reach
LIP maps priority, VIP's the space you breach
MIP controls your reaching arm, AIP will shape your grip
Coordinate transformation, that's the parietal trip
[Verse 2]
LIP creates a salience map, decides where next to gaze
Not just what the stimulus shows, but what deserves your mental maze
Behavioral relevance guides the neurons firing bright
Planning every saccade with spatial attention's might
[Chorus]
PPC's the bridge between, where you look and what you reach
LIP maps priority, VIP's the space you breach
MIP controls your reaching arm, AIP will shape your grip
Coordinate transformation, that's the parietal trip
[Verse 3]
VIP handles near-body space, multisensory and keen
Calculating where you're heading through the visual scene
MIP and parietal reach region move your arm with grace
While AIP reads object shape, configures hand's embrace
[Bridge]
Andersen found the gain fields, neurons that combine
Retinal position signals with body's spatial design
Eye and head position data merge in neural computation
Turning "where on retina" into motor coordination
[Verse 4]
When MT region suffers damage, motion blindness takes its toll
Patient LM cannot see movement, akinetopsia's cruel control
Optic ataxia follows lesions, reaching becomes impaired
Hemispatial neglect emerges when right parietal's been snared
[Verse 5]
Mirror neurons fire when watching others reach and grasp
Social learning through imitation, movement patterns that we clasp
Plasticity allows recovery when stroke has done its harm
Physical therapy rewires circuits, teaching brain to guide the arm
[Chorus]
PPC's the bridge between, where you look and what you reach
LIP maps priority, VIP's the space you breach
MIP controls your reaching arm, AIP will shape your grip
Coordinate transformation, that's the parietal trip
[Outro]
From Newsome's stimulation studies to the fMRI's glow
The dorsal stream's deep secret is the spatial math we know
[Final Chorus]
PPC's the bridge between, where you look and what you reach
LIP maps priority, VIP's the space you breach
MIP controls your reaching arm, AIP will shape your grip
Coordinate transformation, that's the parietal trip
31. The wiring asymmetry
[Verse 1]
Down in the visual cortex maze
Where neurons dance in layered ways
Felleman mapped the wiring game
In macaque brains, the rules the same
Feedforward flows from layers two and three
Straight down to four, like water to the sea
But feedback takes a different route
From five and six, a different suit
[Chorus]
It's not a ladder climbing high
It's not a one-way ticket ride
The feedback outnumbers the feed
A loop that settles, that's what we need
Two-three to four goes up the chain
Five-six comes back like driving rain
More synapses flowing down than up
Vision's not a bottom cup
[Verse 2]
The laminar signature tells the tale
Deep layers send their backwards mail
Avoiding four like toxic ground
In layers one and six they're found
Hierarchical levels get assigned
By this pattern anatomists find
But here's the twist that shakes the tree
Backwards traffic runs more free
[Chorus]
It's not a ladder climbing high
It's not a one-way ticket ride
The feedback outnumbers the feed
A loop that settles, that's what we need
Two-three to four goes up the chain
Five-six comes back like driving rain
More synapses flowing down than up
Vision's not a bottom cup
[Bridge]
V1 gets more from cortical friends
Than from LGN where the signal begins
Higher areas project massively back
Like supervisors checking the stack
Local recurrence joins the choir
Feedback sets the brain on fire
[Verse 3]
Forget the sweep through layered decks
This ain't no simple reflex
The brain's a settlement machine
Where top-down meets the visual scene
Asymmetric wiring rules the day
More connections flow the backwards way
The quantitative truth cuts deep
Vision's not a one-way sweep
[Verse 4]
From MT back to V1's door
The feedback highways carry more
Than forward streams could ever hold
A story that needs to be told
Predictive coding sets the stage
Where expectation writes the page
The brain predicts what it should see
Then checks against reality
[Chorus]
It's not a ladder climbing high
It's not a one-way ticket ride
The feedback outnumbers the feed
A loop that settles, that's what we need
Two-three to four goes up the chain
Five-six comes back like driving rain
More synapses flowing down than up
Vision's not a bottom cup
[Outro]
When you see the world around
Remember how the wires are wound
Backwards flow exceeds the climb
In the brain's asymmetric rhyme
[Final Chorus]
It's not a ladder climbing high
It's not a one-way ticket ride
The feedback outnumbers the feed
A loop that settles, that's what we need
32. What the feedback does (the leading hypotheses)
[Verse 1]
Deep in visual cortex where the neurons fire and glow
Feedback signals streaming from the regions up above
Like a dimmer switch adjusting what the lower layers show
Attention turns the volume up on features that we love
Moran and Desimone discovered how the gain gets turned
When focus hits V4, those neurons bright and clear
While LIP creates a priority map that's finely tuned
Steering early processing to what we need to hear
[Chorus]
Feedback's flowing downward like a river made of thought
Predictions, gain, and context mixing everything we've got
The brain's not just receiving, it's predicting what will be
Error signals climbing up the visual hierarchy
Top-down meets bottom-up in this neural symphony
That's how we see the world through feedback's mystery
[Verse 2]
Your classical receptive field's just part of the whole tale
Surround suppression kicks in from regions far away
An edge responds differently when context tips the scale
Figure-ground and border ownership change throughout the day
Extra-classical influence from feedback connections wide
The same local feature gets a different neural vote
Depending on the global scene that's painted all around
Long-range lateral circuits help the story that it wrote
[Chorus]
Feedback's flowing downward like a river made of thought
Predictions, gain, and context mixing everything we've got
The brain's not just receiving, it's predicting what will be
Error signals climbing up the visual hierarchy
Top-down meets bottom-up in this neural symphony
That's how we see the world through feedback's mystery
[Verse 3]
Rao and Ballard had a vision back in ninety-nine
Predictive coding theory turned the whole field on its head
Higher areas send predictions down the processing line
Lower regions compute the error, send surprises up instead
When prediction matches input, then the error signal fades
Expected edges get explained away and responses shrink
End-stopping and suppression fall in place like puzzle pieces
The iterative dance continues faster than you think
[Bridge]
Perception is construction, not a passive camera's view
The brain's best explanation of the sensory array
Hierarchical Bayesian frameworks try to see it through
Friston's free energy principle maps the broader way
From visual cortex outward to the whole brain's grand design
Though theory runs ahead of what the data can confirm
[Final Chorus]
Feedback's flowing downward like a river made of thought
Predictions, gain, and context mixing everything we've got
The brain's not just receiving, it's predicting what will be
Error signals climbing up the visual hierarchy
Minimizing surprise through this computational spree
That's how we see the world through feedback's mystery
[Outro]
Three hypotheses dancing in the neural night
Attention, context, prediction making vision bright
33. The frequency-band signature of directionality
[Verse 1]
Deep in the cortex where signals collide
Scientists mapped out the neural divide
Bastos and Fries with their laminar probe
Found frequency codes in each brain episode
Feedforward messages ride the fast lane
Gamma waves pulsing through cortical terrain
Forty to eighty hertz climbing up high
From superficial layers reaching the sky
[Chorus]
Alpha beta flowing down the deep layer highway
Gamma theta climbing up the surface byway
Eight to thirty hertz for feedback's call
Forty plus for feedforward through it all
Different frequencies for different directions
Neural multiplexing through oscillation connections
[Verse 2]
Bottom-up evidence surfs on gamma
Theta waves join in this neural drama
Superficial layers broadcast the news
While deep layer channels sing feedback blues
Alpha and beta descend from above
Top-down predictions like hand in glove
Same anatomical loop carries both streams
Separated by frequency engineering schemes
[Chorus]
Alpha beta flowing down the deep layer highway
Gamma theta climbing up the surface byway
Eight to thirty hertz for feedback's call
Forty plus for feedforward through it all
Different frequencies for different directions
Neural multiplexing through oscillation connections
[Bridge]
Communication through coherence speaks the truth
Phase alignment makes the connection proof
When excitability rhythms sync in time
Information flows in perfect rhyme
Granger causality revealed the code
Directional signatures on each neural road
[Verse 3]
Cortical traffic has a measurable sign
Frequency bands drawing the dividing line
Feedforward gamma racing toward the peak
Feedback alpha beta playing hide and seek
Modern neuroscience cracked the mystery wide
How the brain sorts signals on each neural side
Oscillatory channels keep the data clean
Most elegant system you've ever seen
[Verse 4]
Layer four receives the sensory data
Layer two three sends it up like strata
Layer five six connecting back below
Hierarchical processing in rhythmic flow
Visual cortex shows the clearest trace
Each frequency finding its proper place
Million years of evolution's design
Built this multiplexed communication line
[Chorus]
Alpha beta flowing down the deep layer highway
Gamma theta climbing up the surface byway
Eight to thirty hertz for feedback's call
Forty plus for feedforward through it all
Different frequencies for different directions
Neural multiplexing through oscillation connections
[Outro]
In the cortical blues of image processing
Frequency bands keep the signals progressing
From superficial peaks to the deep layer floor
Brain's got the rhythm we're still exploring
34. Why a feedforward sweep still matters
[Verse 1]
In your brain there's a story unfolding fast
When your eyes catch a glimpse of the world so vast
Thorpe discovered the timeline, one fifty milliseconds flat
Feedforward sweep delivers where the action's at
Neural highways climbing up the hierarchy
One quick pass to solve the visual mystery
[Chorus]
Feedforward gets the first guess, coarse but true
Recurrence cleans it up, makes the picture new
Fast then slow, rough then refined
First sweep matters in the processing mind
Don't dismiss that initial flow
It's the foundation that helps you know
[Verse 2]
Through the layers information cascades up
From edges, shapes to objects filling the cup
Ventral stream processing, bottom to the top
That first sweep gives you something before it stops
Rapid categorization, barely time to blink
But enough data flowing to help you think
[Chorus]
Feedforward gets the first guess, coarse but true
Recurrence cleans it up, makes the picture new
Fast then slow, rough then refined
First sweep matters in the processing mind
Don't dismiss that initial flow
It's the foundation that helps you know
[Bridge]
Kar and DiCarlo found the smoking gun
Images that trip up networks when they run
Are the exact same ones where IT takes more time
Biological proof in the paradigm
When feedforward struggles, recurrence kicks in
Extra loops spinning where the problems begin
[Verse 3]
Ambiguity lurking in cluttered scenes
Occlusion hiding what the first pass means
Binding scattered features into unified wholes
Recurrent refinement achieves these goals
But without that rapid initial sweep
The foundation crumbles, processing runs deep
[Verse 4]
Monkey studies with electrodes placed just right
Recording spikes that fire in the visual sight
Early response matches that feedforward flow
Later modulation, that's recurrence you know
Timing tells the story of two-stage design
Fast approximation, then the slow refine
[Chorus]
Feedforward gets the first guess, coarse but true
Recurrence cleans it up, makes the picture new
Fast then slow, rough then refined
First sweep matters in the processing mind
Don't dismiss that initial flow
It's the foundation that helps you know
[Outro]
So remember when you're studying the visual brain
Both pathways working, sunshine and rain
Fast feedforward, then recurrent loops
Two-stage processing, that's how vision groups
First guess matters, refinement too
Together they build the world you view
[Chorus]
Feedforward gets the first guess, coarse but true
Recurrence cleans it up, makes the picture new
Fast then slow, rough then refined
First sweep matters in the processing mind
Don't dismiss that initial flow
It's the foundation that helps you know
35. Within a neuron and across a synapse
[Verse 1]
In the retina where vision starts its dance
Photoreceptors catch each ray of light
No spikes here, just graded potentials' stance
Analog whispers in the fading night
Bipolar cells and horizontal too
Signal with voltage that flows smooth and clean
Short distance chatter, continuous and true
The eye's front stage sets up the visual scene
[Chorus]
From the ganglion forward, spikes take the wheel
All-or-none messages down the neural line
Action potentials make the long haul real
Digital pulses in a spike train design
Glutamate excites, GABA holds it back
Chemical bridges where the neurons meet
Neuromodulators change the neural track
Global state signals make the circuit complete
[Verse 2]
Once ganglion cells fire their electric song
Action potentials race through axon wire
No more graded signals rolling along
Just binary bursts of neural fire
Chemical synapses bridge the gap between
Spikes trigger vesicles to dump their load
Neurotransmitters cross the space unseen
Messages jumping down the neural road
[Chorus]
From the ganglion forward, spikes take the wheel
All-or-none messages down the neural line
Action potentials make the long haul real
Digital pulses in a spike train design
Glutamate excites, GABA holds it back
Chemical bridges where the neurons meet
Neuromodulators change the neural track
Global state signals make the circuit complete
[Verse 3]
Glutamate's the workhorse pushing things ahead
Cortico-cortical highways run on this fuel
AMPA receptors carry signals widespread
NMDA detects coincidence like a jewel
Plasticity and sustained activity flow
Through NMDA's coincidence detection gate
Learning and memory begin to grow
When timing and voltage properly relate
[Verse 4]
Visual cortex layers process what we see
Layer four receives the thalamic drive
Pyramidal cells in two and three
Send feedback signals to keep loops alive
Layer five projects to distant brain parts
While layer six controls the thalamic gate
Each cortical column plays its neural parts
In the grand design of visual fate
[Bridge]
GABA interneurons keep the balance tight
Normalization and gain control their game
Surround suppression sharpens up the sight
Gamma rhythms pulse with inhibitory flame
PV interneurons set the gamma pace
While ACh and noradrenaline set the gain
Arousal and attention find their place
Context switches how the circuits train
[Chorus]
From the ganglion forward, spikes take the wheel
All-or-none messages down the neural line
Action potentials make the long haul real
Digital pulses in a spike train design
Glutamate excites, GABA holds it back
Chemical bridges where the neurons meet
Neuromodulators change the neural track
Global state signals make the circuit complete
[Outro]
From graded potentials to the spike train's beat
Chemical transmission makes it all complete
Excitation, inhibition, modulation's art
Neural symphonies of the visual part
36. The neural code: what the spikes mean
[Verse 1]
When neurons fire their electric song
Action potentials racing along
But what do all these spikes convey?
Four different codes light up the way
Rate code counts them in a window frame
Higher frequency means stronger claim
Contrast and orientation's dance
Written in the firing's advance
[Chorus]
Rate tells the story, population paints the scene
Timing holds secrets, synchrony's the link between
Spikes carry messages in patterns we decode
Four ways to read the neural code
[Verse 2]
Population vectors spread the load
Many neurons share the coding mode
Direction flows through MT's crowd
V1 orientation sings out loud
IT neurons whisper identity
V4 broadcasts color's melody
No single cell can tell the tale
Together they will never fail
[Chorus]
Rate tells the story, population paints the scene
Timing holds secrets, synchrony's the link between
Spikes carry messages in patterns we decode
Four ways to read the neural code
[Bridge]
First spike latency beats the clock
Stronger signals break the block
Milliseconds matter when you need speed
Temporal precision plants the seed
Phase and rhythm lock in time
Binding features, the controversial rhyme
Communication through coherent waves
While the rate code safely saves
[Verse 3]
Fast vision needs that temporal edge
First-spike timing makes the pledge
Contrast encoded in the race
Earlier spikes show stimulus grace
Oscillations mark the beat
When different regions need to meet
Synchrony's the binding glue
Though scientists still argue through
[Verse 4]
From retina to cortex deep
Different layers different beats they keep
Alpha waves and gamma bands
Coordinate across the neural lands
Motor cortex plans the move
Sensory systems set the groove
Hierarchical processing flows
As information upward goes
[Chorus]
Rate tells the story, population paints the scene
Timing holds secrets, synchrony's the link between
Spikes carry messages in patterns we decode
Four ways to read the neural code
[Outro]
From firing rates to phase alignment
Each code serves its assignment
In the brain's electric choir
Multiple meanings never tire
Population, rate, and time
Synchrony completes the rhyme
[Final Chorus]
Rate tells the story, population paints the scene
Timing holds secrets, synchrony's the link between
Spikes carry messages in patterns we decode
Four ways to read the neural code
37. The directional message types (functional summary)
[Verse 1]
In the visual cortex where neurons dance and play
Four directions carry messages in their own special way
Feedforward climbs the ladder with evidence to share
From superficial layers to layer four with care
[Chorus]
Up the chain goes feedforward, "Here's what I cannot predict"
Down comes feedback whispering, "Context makes it click"
Lateral spreads the chatter, normalizing the scene
While neuromodulation sets the global machine
Gamma drives upward, alpha flows back down
Four pathways painting pictures all around
[Verse 2]
Deep layers send their wisdom flowing backward through the stream
Alpha-beta frequencies carrying prediction's dream
Not driving but modulating, biasing what we see
"Attend here" says the feedback, "This is how it ought to be"
[Chorus]
Up the chain goes feedforward, "Here's what I cannot predict"
Down comes feedback whispering, "Context makes it click"
Lateral spreads the chatter, normalizing the scene
While neuromodulation sets the global machine
Gamma drives upward, alpha flows back down
Four pathways painting pictures all around
[Bridge]
GABA keeps the peace within each retinotopic zone
Competition and integration, never works alone
Short-range glutamate connections weave the local tale
While brainstem broadcasts slowly, diffuse without fail
[Verse 3]
From basal forebrain spreading like morning mist so wide
Neuromodulators whisper to every neural side
Arousal, gain, and plasticity, no spatial address
Global states of consciousness in chemical caress
[Verse 4]
Predictive coding algorithms dance in feedback streams
Top-down expectations meeting bottom-up dreams
Error signals propagate when predictions fall short
While lateral inhibition keeps responses in court
[Chorus]
Up the chain goes feedforward, "Here's what I cannot predict"
Down comes feedback whispering, "Context makes it click"
Lateral spreads the chatter, normalizing the scene
While neuromodulation sets the global machine
Gamma drives upward, alpha flows back down
Four pathways painting pictures all around
[Outro]
Layer one and six receive the feedback's gentle touch
Layer four gets driven by the feedforward's clutch
Four messages converging in the visual processing blues
Each pathway tells its story in the neural pathway news
[Final Chorus]
Up the chain goes feedforward, "Here's what I cannot predict"
Down comes feedback whispering, "Context makes it click"
Lateral spreads the chatter, normalizing the scene
While neuromodulation sets the global machine
Gamma drives upward, alpha flows back down
Four pathways painting pictures all around
38. Concrete examples of part-to-part messages
[Verse 1]
From retina ganglion cells the story begins
Twenty channel types encoding what the eye wins
ON-parasol transient contrast in spike trains so neat
Retinotopic addressing, monocular beat
Racing through the optic nerve like messages in code
Each pulse carries meaning down that neural road
[Chorus]
Part to part the signals flow
Through the visual pathways we know
Spike trains dancing, channels singing
Every synapse keeps the message bringing
From RGC to cortex high
That's how we learn to see the sky
[Verse 2]
LGN receives those spikes but adds a twist
Ninety percent non-retinal, can't be missed
Attention gates the relay, arousal sets the tone
Center-surround still intact but no longer alone
Magno streams to four-C-alpha fast and bright
Parvo feeds four-C-beta, contrast just right
[Chorus]
Part to part the signals flow
Through the visual pathways we know
Spike trains dancing, channels singing
Every synapse keeps the message bringing
From RGC to cortex high
That's how we learn to see the sky
[Verse 3]
Layer six sends massive feedback, sharpening the relay
Gain modulation tuning what LGN has to say
Two-three to V-two layer four, edges now appear
Oriented disparity signals crystal clear
Feedforward representation of the visual scene
The sharpest edge detection you have ever seen
[Bridge]
V-four whispers back predictions of shape and space
Attentional gain control keeps the visual pace
MT populations fire for motion direction true
Speed calculations streaming to oculomotor crew
Newsome readout driving saccades and pursuit
Global motion signals bearing perceptual fruit
[Verse 4]
IT sends identity signals, invariant and strong
To perirhinal cortex where memories belong
Prefrontal gets the message for decisions to make
Patch-clamp electrodes show what pathways take
GABA glutamate dancing, receptors in their place
Gamma up and beta down, the spectral embrace
[Verse 5]
Dorsal stream processes where objects might be found
While ventral stream determines if they're safe and sound
Parietal integration maps the spatial array
Temporal lobe semantics name what's on display
Action potentials coding every shade and hue
Building conscious vision from electric blue
[Chorus]
Part to part the signals flow
Through the visual pathways we know
Spike trains dancing, channels singing
Every synapse keeps the message bringing
From RGC to cortex high
That's how we learn to see the sky
[Outro]
Population codes revealing secrets of the mind
Temporal patterns showing what we're meant to find
Each connection tells a story of the light we see
Part to part forever, visual symphony
39. Lesion and ablation studies (necessity)
[Verse 1]
When Ungerleider took his blade to monkey minds
Cut the temporal stream, left the parietal behind
What pathway broke, what pathway stayed intact?
Remove a piece, observe the deficit exact
Schiller split the magno from the parvo layers clean
Motion died but color lived, proved channels aren't just seen
[Chorus]
Cut it out, watch what breaks
That's the method old science makes
Lesion tells you what's required
Not the where, just what's desired
Necessity's the only claim
When you're playing the ablation game
[Verse 2]
Nature throws her own destructive dice around
Stroke in V4, chromatic world shuts down
Achromatopsia steals the rainbow's gift
MT damage leaves you motion parallax drift
Akinetopsia freezes life to still frames
While fusiform lesions steal familiar faces' names
[Chorus]
Cut it out, watch what breaks
That's the method old science makes
Lesion tells you what's required
Not the where, just what's desired
Necessity's the only claim
When you're playing the ablation game
[Bridge]
But here's the trap that catches eager minds
The area you cut might just be roads and signs
Could be input, could be relay station
Not the place of computation
Double dissociation's the golden key
D.F. and optic ataxia set the circuits free
[Verse 3]
V1 vanishes, conscious sight goes black
But blindsight proves unconscious processing tracks
Prosopagnosia leaves you face-blind and confused
Fusiform's necessity clearly well-perused
Chemical or surgical, monkey or human brain
The oldest method still breaks the cognitive chain
[Verse 4]
Transcranial magnets pulse and interfere
Virtual lesions making neural functions disappear
Millisecond timing shows the causal flow
When disruption stops the cognitive show
Pharmacological blocks at the synapse divide
Show which transmitters let the signals ride
[Chorus]
Cut it out, watch what breaks
That's the method old science makes
Lesion tells you what's required
Not the where, just what's desired
Necessity's the only claim
When you're playing the ablation game
[Outro]
From laboratory table to the trauma ward
Lesion studies keep revealing what the brain's stored
Function follows damage in this ancient dance
Remove the part, lose the chance
[Final Chorus]
Cut it out, watch what breaks
That's the method old science makes
Lesion tells you what's required
Not the where, just what's desired
Necessity's the only claim
When you're playing the ablation game
40. Single-unit and multi-unit recording (correlation, fine grain)
[Verse 1]
Slide that microelectrode deep inside
Listen to the whispers neurons can't hide
Hubel and Wiesel with their steady hand
Built the foundation we still understand
V1 columns arranged in perfect rows
Simple cells fire when the line just flows
Complex cells dancing to orientation's call
Nobel prize waiting at the end of it all
[Chorus]
Record and correlate, that's how we see
What makes each neuron sing in harmony
From one electrode to a thousand strong
Population decoding proves the brain's song
But correlation ain't causation's twin
The bystander problem lurks within
[Verse 2]
Extended reaching to every domain
MT neurons love direction's refrain
V4 curved lines make the circuits glow
IT cortex where the objects show
Face patches mapping every subtle feature
Each electrode becomes our neural teacher
DiCarlo's decoding breaks the mystery wide
Information flows where behaviors reside
[Chorus]
Record and correlate, that's how we see
What makes each neuron sing in harmony
From one electrode to a thousand strong
Population decoding proves the brain's song
But correlation ain't causation's twin
The bystander problem lurks within
[Bridge]
Neuropixels dense arrays pierce the veil
Hundreds firing, telling neural tales
Linear readouts from the cortical stage
Unlocking secrets of the visual age
Single units whisper their tuning curves
Multi-unit chorus shows what each preserves
Fine grain analysis reveals the code
But causation needs another road
[Verse 3]
Ocular dominance carved in living stone
Simple complex cells each in their zone
Microelectrodes trace the visual stream
From retina's light to perception's dream
The method that built our field stands tall
Recording spikes that rise and fall
Correlation maps the neural terrain
But proving cause needs a different game
[Verse 4]
Spike trains dancing through the neural maze
Temporal patterns in millisecond phase
Rate codes whisper what the neurons know
While timing codes add to the show
Signal-to-noise ratios guide our way
Through cortical circuits night and day
Each action potential tells its tale
In the language where perceptions never fail
[Final Chorus]
Record and correlate, that's how we see
What makes each neuron sing in harmony
From one electrode to a thousand strong
Population decoding proves the brain's song
But correlation ain't causation's twin
The bystander waits to catch us in
[Outro]
Single units tell their faithful story
Multi-unit arrays reveal the glory
Fine grain recordings light the neural way
Building knowledge spike by spike each day
41. Microstimulation (sufficiency, the causal complement)
[Verse 1]
Deep in the cortex where neurons fire and dance
Scientists wondered if patterns meant more than chance
They saw correlations in electrical waves
But correlation's not causation, that's what science craves
So they took their tiny electrode, made it dual-purpose tool
Record the signals, then stimulate - that's the golden rule
[Chorus]
Microstimulation, sufficiency's the game
Push the local circuits, see if behavior's changed
From correlation to causation, that's the leap we need
When you drive the neurons, does it plant the perceptual seed?
Sufficiency proven when the current makes it so
The causal complement that helps our knowledge grow
[Verse 2]
Newsome found the magic in area MT
Direction columns dancing with their speciality
First he recorded while the monkey watched dots move
Neurons fired strongest when their preference was proved
Then he flipped the switch, sent current through the same spot
Suddenly the monkey's judgment shifted quite a lot
[Chorus]
Microstimulation, sufficiency's the game
Push the local circuits, see if behavior's changed
From correlation to causation, that's the leap we need
When you drive the neurons, does it plant the perceptual seed?
Sufficiency proven when the current makes it so
The causal complement that helps our knowledge grow
[Verse 3]
Tsao took the concept to the face-selective patch
Stimulated gently, waited for perception's catch
The macaque's vision twisted toward familiar form
Even random patterns looked like faces in the storm
These experiments together paint the clearest view
That neural activity doesn't just correlate, it drives what's true
[Bridge]
Recording shows the correlation strong and clear
But stimulation proves what scientists hold dear
The productive direction of the causal chain
When we activate the area, perception feels the strain
Sufficiency established by the current's gentle touch
Proving neural populations control our senses much
[Chorus]
Microstimulation, sufficiency's the game
Push the local circuits, see if behavior's changed
From correlation to causation, that's the leap we need
When you drive the neurons, does it plant the perceptual seed?
Sufficiency proven when the current makes it so
The causal complement that helps our knowledge grow
[Outro]
So remember when you study how the brain constructs our sight
Recording plus stimulation makes the causal picture bright
The textbook proof of causation lives in these experiments bold
Where tiny currents unlock secrets that the neurons hold
42. Optogenetics (causal, cell-type-specific, millisecond-precise)
[Verse 1]
In the neuroscience lab where precision reigns supreme
Scientists discovered something beyond their wildest dreams
Channelrhodopsin from algae, responds to flashes of blue
While halorhodopsin silences when yellow light breaks through
Engineer these proteins into neurons of your choice
Give each cell type a different photosensitive voice
[Chorus]
Optogenetics, the causal scalpel so refined
Millisecond timing with cell-type control combined
Flash the light to excite, switch the hue to make it quiet
No more guessing at connections, now we've got the power to try it
Channelrhodopsin excites, halorhodopsin inhibits
Gold standard for causation when you need those clean exhibits
[Verse 2]
Target just the PV interneurons, leave the pyramidal alone
Unlike crude electrical stimulation, precision's finally shown
Silence V1 with a laser pulse, downstream responses fade away
Behavior crumbles in an instant, proving vision's neural highway
Reversible manipulation, switch it on then switch it off
What seemed impossible before now seems remarkably soft
[Chorus]
Optogenetics, the causal scalpel so refined
Millisecond timing with cell-type control combined
Flash the light to excite, switch the hue to make it quiet
No more guessing at connections, now we've got the power to try it
Channelrhodopsin excites, halorhodopsin inhibits
Gold standard for causation when you need those clean exhibits
[Verse 3]
Inhibitory circuits generate gamma rhythms in the brain
Different interneuron classes each control a different domain
Recurrent loops, not feedforward, necessary for detection tasks
Light reveals the hidden circuits, answers questions no one asks
Mice respond with perfect timing, primates prove more challenging still
But advancing techniques promise we'll soon bend them to our will
[Verse 4]
Fiber optics thread through tissue, delivering photons precise
Viral vectors carry genes that make the neurons think twice
Blue light opens sodium gates, yellow shuts them down with ease
Temporal resolution matched only by the mind's unease
Fear conditioning and memory, social behavior too
Every circuit mapped and tested, every pathway coming through
[Bridge]
From correlation to causation, that's the scientific dream
Necessity and sufficiency revealed in laser's beam
Cell-type-resolved precision cuts through neural network maze
Optogenetics revolutionized our experimental ways
[Chorus]
Optogenetics, the causal scalpel so refined
Millisecond timing with cell-type control combined
Flash the light to excite, switch the hue to make it quiet
No more guessing at connections, now we've got the power to try it
Channelrhodopsin excites, halorhodopsin inhibits
Gold standard for causation when you need those clean exhibits
[Outro]
When you need to prove that neural circuit does the job
Optogenetics is the tool that makes the neurons throb
43. fMRI (whole-brain, correlational, coarse spatial, human)
[Verse 1]
In the scanner's magnetic embrace so strong
Blood oxygen tells the neural song
BOLD signals dancing through your brain
Millimeter maps where visions reign
Seconds tick by while we decode
The visual pathways, neuron's abode
Non-invasive window to the mind
Whole-brain secrets waiting to find
[Chorus]
BOLD and bright, the signal flows
Where the visual cortex glows
Correlations paint the scene
But causation stays unseen
Coarse resolution tells the tale
Of how human visions never fail
BOLD and bright across V1
Till the visual processing's done
[Verse 2]
Traveling waves sweep 'cross the screen
Mapping borders sharp and clean
V1, V2, V3 appear
Retinotopic boundaries crystal clear
Cortical magnification shown
In living flesh and blood and bone
Each visual area finds its place
Through stimulus patterns we can trace
[Chorus]
BOLD and bright, the signal flows
Where the visual cortex glows
Correlations paint the scene
But causation stays unseen
Coarse resolution tells the tale
Of how human visions never fail
BOLD and bright across V1
Till the visual processing's done
[Verse 3]
Functional localizers stake their claim
Faces over objects play the game
FFA emerges from the noise
While PPA gives scenes its voice
Motion beats static every time
To find hMT+ in neural rhyme
Category patches come alive
When contrast signals help them thrive
[Verse 4]
MVPA reads the voxel code
Patterns dancing down the road
Classifiers decode what you see
From neural activity running free
Early cortex holds the key
To reconstruct what eyes perceive
Rough images from brain waves drawn
The workhorse method carries on
[Bridge]
But remember what this method lacks
Hundreds of thousands in each tract
Voxels pool neurons by the score
BOLD runs slow, can't give us more
No single cells beneath our view
No causation breaking through
Just correlations, rich and wide
Across the brain's visual divide
[Verse 5]
Network analysis reveals the flow
How visual regions come and go
Connectivity maps the brain's design
Where feedforward meets the feedback line
Laminar patterns start to show
Which layers make the signals grow
Top-down predictions meet their mark
In this neural feedback arc
[Chorus]
BOLD and bright, the signal flows
Where the visual cortex glows
Correlations paint the scene
But causation stays unseen
Coarse resolution tells the tale
Of how human visions never fail
BOLD and bright across V1
Till the visual processing's done
[Outro]
In the blues of brain scanning light
BOLD signals burning bright
Human vision's secrets told
Through the story BOLD unfolds
44. TMS (causal, human, coarse)
[Verse 1]
Doctor holds a coil above your head
Magnetic pulses through your skull instead
Of cutting brain tissue like before
We zap the cortex through an open door
No surgery needed, just a gentle hum
Creating virtual lesions, here they come
Transcranial stimulation, that's the name
Disrupting neurons in this research game
[Chorus]
TMS shows us causal links so clear
When motion vanishes, hMT's the gear
V1 gets zapped, phosphenes appear
Chronometry tells us when circuits steer
Transient disruption, then it's gone
Causal necessity, the proof lives on
[Verse 2]
Point the coil at visual area one
Flickering lights dance, the show's begun
Phosphenes sparkle in your visual field
Masking stimuli that once were revealed
Time the pulses just before you see
We'll map the moment when V1 runs free
Surface cortex only, can't go deep
Centimeter patches, that's what we reap
[Chorus]
TMS shows us causal links so clear
When motion vanishes, hMT's the gear
V1 gets zapped, phosphenes appear
Chronometry tells us when circuits steer
Transient disruption, then it's gone
Causal necessity, the proof lives on
[Verse 3]
Target hMT plus, watch motion die
Moving dots stand still before your eye
The waterfall effect just disappears
Proving what we've suspected all these years
This patch of cortex makes the world flow
Without it working, motion doesn't show
Human subjects, healthy and intact
Virtual lesions prove the causal fact
[Bridge]
Correlation whispers, causation shouts
TMS removes all reasonable doubts
Spatially coarse but temporally tight
Millisecond timing gets the sequence right
Ethics smile because it's not for keeps
Reversible lesions while the subject sleeps
[Verse 4]
Single pulses or a rapid train
Either way we interrupt the brain
Double dissociation is the goal
Each region playing its specific role
BOLD signals tell us correlation
TMS brings causal demonstration
From the scanner to the magnet chair
We build the map of what happens where
[Chorus]
TMS shows us causal links so clear
When motion vanishes, hMT's the gear
V1 gets zapped, phosphenes appear
Chronometry tells us when circuits steer
Transient disruption, then it's gone
Causal necessity, the proof lives on
[Outro]
Magnetic coil above the visual brain
Proving causation without lasting pain
From correlation to the causal truth
TMS provides the scientific proof
[Final Chorus]
TMS shows us causal links so clear
When motion vanishes, hMT's the gear
V1 gets zapped, phosphenes appear
Chronometry tells us when circuits steer
Transient disruption, then it's gone
Causal necessity, the proof lives on
45. EEG / MEG (timing, human, correlational)
[Verse 1]
Put electrodes on the scalp, watch the brain waves flow
EEG catches electrical storms where neurons glow
While fMRI shows us the where with spatial precision fine
EEG and MEG reveal the when in millisecond time
Magnetic fields dance through the skull, no contact required
MEG sensors floating overhead, neural timing acquired
[Chorus]
Timing staircase climbing up, millisecond by millisecond rise
Population activity bursts before your very eyes
N170 face component peaks at one-seventy time
Representational dynamics show how neural patterns climb
When feedforward meets recurrence in the temporal maze
EEG and MEG decode the brain's millisecond phase
[Verse 2]
First the visual cortex sparks in early waves of sight
Then cascading through the regions like dominoes in flight
Face-selective N170 component marks its spot
One hundred seventy milliseconds, faces trigger what
Neural populations fire in synchronized array
Scalp recordings catch the rhythm of the processing display
[Chorus]
Timing staircase climbing up, millisecond by millisecond rise
Population activity bursts before your very eyes
N170 face component peaks at one-seventy time
Representational dynamics show how neural patterns climb
When feedforward meets recurrence in the temporal maze
EEG and MEG decode the brain's millisecond phase
[Verse 3]
Decoding over time reveals how representations grow
From simple edges into objects as the patterns flow
Feedforward pathways push ahead with bottom-up command
Then recurrent loops kick in with top-down feedback grand
The timing tells the story that spatial maps can't say
When neural processes unfold throughout the processing day
[Verse 4]
Alpha waves and gamma rhythms mark the neural beat
Delta, theta, beta bands make oscillation complete
Event-related potentials rise above the background noise
P300 and MMN give cognition its voice
Evoked responses locked in time to stimulus events
Show how the brain reacts and what each signal represents
[Bridge]
Weak on spatial localization but strong on temporal beat
Complementing fMRI to make the picture complete
Electric storms and magnetic waves reveal the neural dance
While millisecond precision gives timing its chance
[Chorus]
Timing staircase climbing up, millisecond by millisecond rise
Population activity bursts before your very eyes
N170 face component peaks at one-seventy time
Representational dynamics show how neural patterns climb
When feedforward meets recurrence in the temporal maze
EEG and MEG decode the brain's millisecond phase
[Outro]
When you need to know the timing of the neural show
EEG and MEG will let the millisecond secrets flow
46. Connectomics (anatomy, ground truth wiring)
[Verse 1]
Deep beneath the microscope's electric eye
Serial sections cut so thin they barely cry
Electron beams reveal what light cannot
Every synapse mapped, every connection caught
The retina was first to bare its soul
Circuit diagrams made beautifully whole
[Chorus]
Wire by wire, we trace the brain's design
Connectomics shows us every line
From electron scopes to tracers in the stream
Mapping out the visual processing dream
Structure tells us where the signals flow
But not the secrets of what they truly know
[Verse 2]
MICrONS took a cubic millimeter slice
Mouse visual cortex paid the ultimate price
Hundred thousand cells with functional light
Two-photon recordings burning laser bright
Structure plus function in perfect dance
Ground truth wiring gets its only chance
[Chorus]
Wire by wire, we trace the brain's design
Connectomics shows us every line
From electron scopes to tracers in the stream
Mapping out the visual processing dream
Structure tells us where the signals flow
But not the secrets of what they truly know
[Verse 3]
Inject the tracers, anterograde they crawl
Retrograde pathways answer anatomy's call
Felleman and Van Essen drew the map
Hierarchical connections bridge the gap
Markov and Kennedy weighted every thread
Quantified the paths where vision signals spread
[Verse 4]
C. elegans showed us the way to start
Three hundred neurons mapped from every part
Complete connectome of a nervous mind
First full wiring diagram of its kind
But complexity scales beyond our dreams
When mammalian circuits burst the seams
[Bridge]
Diffusion MRI in living human brain
White matter bundles dancing through the rain
Coarse resolution but the streams shine through
What connects to what, we finally knew
Dorsal, ventral pathways carved in white
Tractography reveals the visual flight
[Chorus]
Wire by wire, we trace the brain's design
Connectomics shows us every line
From electron scopes to tracers in the stream
Mapping out the visual processing dream
Structure tells us where the signals flow
But not the secrets of what they truly know
[Chorus]
Wire by wire, we trace the brain's design
Connectomics shows us every line
From electron scopes to tracers in the stream
Mapping out the visual processing dream
Structure tells us where the signals flow
But not the secrets of what they truly know
[Outro]
The wiring diagram's just the start we find
Computation hides behind each neural bind
Connectomics builds the foundation strong
For understanding how perception flows along
47. Two-photon and calcium imaging (correlation, many cells, spatial layout)
[Verse 1]
Deep in cortex, neurons hide their secrets well
But GCaMP proteins break the spell
When calcium floods and action potentials race
Green fluorescence lights up every face
Two-photon lasers pierce the living brain
Thousand cells at once, no tissue strain
[Chorus]
Correlation, many cells, spatial display
Pinwheels spinning in V1's array
Orientation maps come alive in real time
Columnar secrets revealed in fluorescent rhyme
Many cells, correlation, spatial sight
Brain's true architecture burning bright
[Verse 2]
Wide-field imaging sweeps the dorsal scene
Whole cortex dancing on computer screen
Chronic windows let us track the same
Neural populations through learning's game
Single-cell resolution tells the tale
Of how visual processing will never fail
[Chorus]
Correlation, many cells, spatial display
Pinwheels spinning in V1's array
Orientation maps come alive in real time
Columnar secrets revealed in fluorescent rhyme
Many cells, correlation, spatial sight
Brain's true architecture burning bright
[Bridge]
Rodent models show us what we need to know
Optical access limits how deep we can go
Correlational data streams like blues
But spatial organization breaks the news
Genetically encoded, calcium's dance
Gives neural networks their fighting chance
[Verse 3]
Learning changes written in calcium's glow
Same cells tracked as memories grow
Confirmed the columns that we theorized
Made invisible patterns crystallized
From hypothesis to visual proof
Two-photon microscopy lifts the roof
[Verse 4]
Signal-to-noise battles in the microscope's eye
Photon statistics make the data fly
Stimulus-evoked responses shine so clear
While spontaneous activity disappears
Multi-session tracking through weeks and days
Maps the plastic brain's adaptive ways
[Chorus]
Correlation, many cells, spatial display
Pinwheels spinning in V1's array
Orientation maps come alive in real time
Columnar secrets revealed in fluorescent rhyme
Many cells, correlation, spatial sight
Brain's true architecture burning bright
[Outro]
When neurons fire, calcium flows
And GCaMP is how the story goes
Spatial layout, correlation's key
Many cells singing in harmony
[Final Chorus]
Correlation, many cells, spatial display
Brain's true secrets lighting up the way
Many cells, correlation, spatial sight
Neural symphony burning bright
48. Computational model comparison (explanatory adequacy)
[Verse 1]
Scientists built machines that see like we do
Deep networks trained on objects, layer by layer grew
They tested against neurons in macaque brains
Found something shocking in the data chains
V4 and IT cortex responses matched the code
Better than any hand-built model ever showed
[Chorus]
Brain-Score is the benchmark, measuring how well
Your model's representations can predict and tell
What neurons fire when images appear
Sufficiency's the game, but causation's never clear
You can match the data, match behavior too
But that don't mean the brain works just like you
[Verse 2]
Yamins and DiCarlo cracked the mystery wide
Convolutional layers mirror what's inside
From V1 through V2, up to higher ground
Each network layer maps where brain responses are found
CORnet added feedback loops to match the time
When signals flow backward, keeping perfect rhyme
[Chorus]
Brain-Score is the benchmark, measuring how well
Your model's representations can predict and tell
What neurons fire when images appear
Sufficiency's the game, but causation's never clear
You can match the data, match behavior too
But that don't mean the brain works just like you
[Bridge]
Evidence in its own right, these models stake their claim
Computational theories playing the prediction game
What it proves is powerful - your theory could be right
What it cannot prove is mechanism's inner sight
Necessary but not sufficient for the mechanistic case
Multiple roads can lead to the same neural place
[Verse 3]
From pixels to perception, hierarchies unfold
Edge detectors, curves, and shapes, the story's being told
Object recognition emerges from the stack
But implementation details, well, that knowledge we still lack
The transformation's captured, the function's crystal clear
But how the wetware does it remains a frontier
[Verse 4]
ResNet, VGG, and AlexNet compete
Each architecture striving to make predictions complete
ImageNet pretrained then fitted to the brain
Silicon circuits learning biological refrain
Yet questions linger in the depths of neural space
Do brains compute like GPUs in this computational race?
[Final Chorus]
Brain-Score is the benchmark, measuring how well
Your model's representations can predict and tell
What neurons fire when images appear
Sufficiency's the game, but causation's never clear
You can match the data, match behavior too
But the brain's true secrets are still breaking through
49. How the methods combine into proof
[Verse 1]
One experiment's just a whisper in the dark
Recording neurons firing their electric spark
MT cells love motion, that's what electrodes say
But whispers can deceive us, lead our minds astray
[Chorus]
When the methods converge, that's when we know it's true
Five different pathways painting the same view
Recording, lesions, stimulation too
Human studies make it four and five, breakthrough
Convergence is the key, not just one or two
That's how scattered clues become the proof we use
[Verse 2]
Lesion studies cut the circuits, watch what falls apart
Akinetopsia strikes when MT can't restart
But damage spreads like wildfire, hits more than one zone
A single method's verdict leaves us still unknown
[Chorus]
When the methods converge, that's when we know it's true
Five different pathways painting the same view
Recording, lesions, stimulation too
Human studies make it four and five, breakthrough
Convergence is the key, not just one or two
That's how scattered clues become the proof we use
[Verse 3]
Microstimulation whispers lies into the brain
Monkey sees false motion dancing in the rain
TMS disrupts the human visual stream
Each method tells a fragment of a larger dream
[Bridge]
FMRI blood flow maps the territory bright
MT area glowing in the scanner's sight
Five independent witnesses all agree
Motion lives in MT, that's our certainty
[Verse 4]
But where the methods scatter, pointing different ways
The brain map stays uncertain through our searching days
No single golden hammer builds the neural truth
Convergence paints the picture, that's our living proof
[Verse 5]
In the lab at midnight, questions multiply
Single-cell recordings make the neurons cry
Population vectors dancing left and right
But without convergence, we're still blind in flight
Replication crisis haunts the ivory tower
One method standing lonely has no proving power
[Chorus]
When the methods converge, that's when we know it's true
Five different pathways painting the same view
Recording, lesions, stimulation too
Human studies make it four and five, breakthrough
Convergence is the key, not just one or two
That's how scattered clues become the proof we use
[Chorus]
When the methods converge, that's when we know it's true
Five different pathways painting the same view
Recording, lesions, stimulation too
Human studies make it four and five, breakthrough
Convergence is the key, not just one or two
That's how scattered clues become the proof we use
[Outro]
Five roads to one conclusion, that's the scientist's way
Scattered evidence united saves the day
50. Response latency staircase (macaque, foveal, high-contrast)
[Verse 1]
Light hits the retina, photoreceptors wake
Twenty milliseconds for the chemical shake
Brighter means faster through that transduction dance
RGCs fire by fifty, giving vision its chance
Through the optic nerve the signal races on
LGN at thirty-five, relay station strong
[Chorus]
It's the staircase of response, climbing through the brain
Every millisecond counted in the visual chain
From photons to perception, timing tells the tale
V1 to V4 ascending, IT brings up the trail
One hundred fifty total for that first decision made
The hierarchy revealed by each synaptic delay
[Verse 2]
V1 lights up at forty, magno cells lead the way
Parvo trails behind them by a few millisecond's stay
V2 joins at sixty with its complex feature maps
V4 follows after with attention's binding gaps
But here's the plot twist in this neural narrative flow
MT jumps the queue at fifty, stealing V4's show
[Chorus]
It's the staircase of response, climbing through the brain
Every millisecond counted in the visual chain
From photons to perception, timing tells the tale
V1 to V4 ascending, IT brings up the trail
One hundred fifty total for that first decision made
The hierarchy revealed by each synaptic delay
[Verse 3]
The magno superhighway rockets information through
Dorsal stream for action beats the ventral's careful view
While IT contemplates objects from eighty to one-ten
Your reflexes already moved your hand away again
Conscious recognition waits for two hundred or more
By then your brain's decided what this image is for
[Verse 4]
Single cell recordings tell us where and when they fire
Each electrode listening to neural choir
Spike trains mark the moments when the neurons come alive
Millisecond precision shows how visual circuits thrive
From anesthesia's window to the awake and alert state
The timing never lies about each processing gate
[Bridge]
Thorpe discovered something in that ultra-rapid test
One feedforward sweep is all you need to beat the rest
No time for reconsideration, no loops back to start
Just layer after layer playing its neural part
The strongest evidence we have for how the system's built
Is written in the timing, millisecond to the hilt
[Chorus]
It's the staircase of response, climbing through the brain
Every millisecond counted in the visual chain
From photons to perception, timing tells the tale
V1 to V4 ascending, IT brings up the trail
One hundred fifty total for that first decision made
The hierarchy revealed by each synaptic delay
[Outro]
From macaque to human, the pattern stays the same
Response latency staircase, time-stamped neural fame
High contrast and foveal, the conditions crystal clear
The brain's processing order written in what we hear
51. Cell and channel counts (human, per eye unless noted)
[Verse 1]
In your retina's dark cathedral, rods are scattered thick
One hundred twenty million strong, they catch the dimmest flick
But cones are precious jewels, just six million bright
They paint the world in color when there's plenty of light
These photoreceptors dance, converting rays to code
Before they pass their message down the neural road
[Chorus]
One-twenty million rods for darkness, six million cones for hue
Ganglion cells compress it down to one point five million through
The optic nerve's thin cable carries one megabit per second
From eye to brain the data streams, reality's been reckoned
[Verse 2]
Retinal ganglion cells collect this massive flood
Fifteen to twenty different types sort through the visual mud
They squeeze all that information through a bottleneck so tight
One million axons bundled up to carry sound and sight
The optic nerve's bandwidth is surprisingly lean
Just one megabit streaming through this biological machine
[Chorus]
One-twenty million rods for darkness, six million cones for hue
Ganglion cells compress it down to one point five million through
The optic nerve's thin cable carries one megabit per second
From eye to brain the data streams, reality's been reckoned
[Verse 3]
The LGN receives this flood with matching neuron count
One million relay stations that control the data's route
But here's the shocking twist that neuroscience has found
Just five to ten percent from retina comes inbound
The rest is feedback, context, memory's embrace
The brain talks to itself while building visual space
[Verse 4]
Your visual cortex layers stack like city floors
Each one extracts new features from the data that it stores
Simple cells find edges, complex cells track motion
While hypercomplex neurons fuel the visual devotion
Parallel pathways split the stream in fascinating ways
The "what" and "where" of vision through distinct neural arrays
[Bridge]
V1 explodes the signal wide
One hundred fifty million neurons on each side
Could reach two-eighty million strong
Processing edges, motion, getting nothing wrong
Thirty areas in the cortex vie
Half the primate brain just to see with one eye
[Chorus]
One-twenty million rods for darkness, six million cones for hue
Ganglion cells compress it down to one point five million through
From sparse to dense to distributed, the visual pathway flows
How your brain constructs the world, now everybody knows
[Chorus]
One-twenty million rods for darkness, six million cones for hue
Ganglion cells compress it down to one point five million through
The optic nerve's thin cable carries one megabit per second
From eye to brain the data streams, reality's been reckoned
[Outro]
From photons hitting retina to cortex painting scenes
The numbers tell the story of our visual processing dreams
52. Receptive field size progression (rough, at fixed eccentricity)
[Verse 1]
Down in the retina where vision starts to grow
Ganglion cells are tiny, just a tenth degree or so
They catch the smallest details in the foveal core
But as signals climb the ladder, fields expand much more
From RGC to cortex, each stage multiplies by ten
Trading sharp precision for a broader viewing pen
[Chorus]
Ten times bigger, ten times wider
As you climb the neural tower
Point-one grows to one degree
V2 doubles what you see
Four to eight in V4's domain
MT spreads across the plane
IT sees fifty degrees wide
Receptive fields on a climbing ride
[Verse 2]
V1 takes that retinal whisper, makes it half a degree
Simple cells and complex ones processing what we see
But V2 doubles down the bet, two degrees in span
Building textures, shapes, and edges with a broader plan
Each ascending cortical stage sacrifices the precise
For invariance and abstraction, that's the neural price
[Chorus]
Ten times bigger, ten times wider
As you climb the neural tower
Point-one grows to one degree
V2 doubles what you see
Four to eight in V4's domain
MT spreads across the plane
IT sees fifty degrees wide
Receptive fields on a climbing ride
[Bridge]
From arcminutes to degrees, the progression never lies
Motion flows through MT's channels, ten degrees comprise
While IT neurons span the vista, fifty degrees or more
Including foveal territory in their receptive store
Spatial detail sacrificed for categorical might
Abstract recognition blooming from the sharpest sight
[Verse 3]
V4 color processing with its eight-degree embrace
Objects forming, surfaces merging across visual space
The hierarchy's ancient wisdom, magnification's law
Every stage pays resolution for the patterns that it saw
From fine-grained local features to the forest from the trees
Visual cortex builds its empire through expanding boundaries
[Verse 4]
Evolution's clever solution to the processing bind
How to capture both the details and the scene combined
Parallel pathways streaming upward through the visual maze
What and where streams diverging in their separate ways
Small fields catch the features while the large ones track the flow
This distributed architecture makes perception grow
[Chorus]
Ten times bigger, ten times wider
As you climb the neural tower
Point-one grows to one degree
V2 doubles what you see
Four to eight in V4's domain
MT spreads across the plane
IT sees fifty degrees wide
Receptive fields on a climbing ride
[Outro]
So remember when you're learning how the brain decodes the scene
Size progression tells the story of the visual machine
Trading pixels for the meaning, local detail for the whole
Receptive field expansion plays the starring role
53. Cortical magnification
[Verse 1]
Deep inside your visual cortex, there's a curious design
Where the center gets the spotlight, and the edges fall behind
Eight degrees of central vision take up half of V1's space
While the periphery gets crowded in a much more narrow place
[Chorus]
Magnification, magnification
Fovea gets the royal treatment in your brain
More millimeters per degree, that's the cortical decree
Why your sharpest sight lives in the center lane
Magnification, magnification
[Verse 2]
From the fovea packed with cone cells to the midget ganglion chain
Every photon gets more real estate when it reaches visual brain
But move away from center stage, watch the cortex space decline
Inverse with eccentricity, following that simple line
[Chorus]
Magnification, magnification
Fovea gets the royal treatment in your brain
More millimeters per degree, that's the cortical decree
Why your sharpest sight lives in the center lane
Magnification, magnification
[Bridge]
Try to read these very words while looking to the side
Peripheral acuity has nowhere left to hide
Dense-packed cones need dense-packed cortex, that's the neural rule
Evolution built this bias as a survival tool
[Verse 3]
When you fixate on a target, know your brain's unequal bet
Half its visual processing power on ten degrees is set
Central vision gets the mansion, edges get the studio flat
Cortical magnification explains exactly where you're at
[Verse 4]
In the monkey studies showing this anatomical divide
Scientists mapped the neurons where your visual signals hide
Each square millimeter matters in this precious neural land
Central vision claims the territory that the cortex has on hand
[Chorus]
Magnification, magnification
Fovea gets the royal treatment in your brain
More millimeters per degree, that's the cortical decree
Why your sharpest sight lives in the center lane
Magnification, magnification
[Bridge 2]
Saccadic eye movements dance to keep the target centered right
Your brain knows where the processing power gives the clearest sight
The foveal advantage isn't just about the photoreceptors there
It's magnified again when cortical neurons do their share
[Chorus]
Magnification, magnification
Fovea gets the royal treatment in your brain
More millimeters per degree, that's the cortical decree
Why your sharpest sight lives in the center lane
Magnification, magnification
[Outro]
From retinal density to cortical destiny
The center always wins the magnification game
54. The motifs artificial networks took from biology
[Verse 1]
Back in nineteen-eighty Fukushima saw the light
Hubel and Wiesel's simple cells caught his sight
He built the Neocognitron, layer upon layer
Convolution meets pooling, nature's neural player
Filter then pool, that's the hierarchy's call
From retina's blueprint to the artificial hall
[Chorus]
Biology's whispers in silicon dreams
Convolution and pooling, simple and complex schemes
From edges to textures, parts to the whole
Gabor detectors emerge from the code
The brain's ancient wisdom in networks reborn
Teaching machines how vision is worn
[Verse 2]
First layer convergence, a remarkable sight
Gabor-like filters learning edge detection's might
Oriented lines appear without human design
The same family V1 uses, crossing that line
Natural images train what evolution knew
Artificial features matching biology's view
[Chorus]
Biology's whispers in silicon dreams
Convolution and pooling, simple and complex schemes
From edges to textures, parts to the whole
Gabor detectors emerge from the code
The brain's ancient wisdom in networks reborn
Teaching machines how vision is worn
[Verse 3]
Hierarchy climbs from V1 to IT's peak
V2, V4 stages that neural networks seek
Edges become textures, textures turn to parts
Parts build up objects, artificial arts
The correspondence loose but measurable and real
Abstraction's ladder that both systems feel
[Bridge]
Divisive normalization found its digital twin
Batch norm and layer norm, letting gradients win
Attention mechanisms weight by relevance computed
Gain modulation's concept in transformers rooted
Center-surround whitening from retina's DoG
Input decorrelation, breaking visual fog
[Verse 4]
Difference of Gaussians cleans the visual stream
Whitening preprocessing, the retina's scheme
Surround suppression keeps the signal pure
Contrast normalization, cortex signature
Every innovation has its biological thread
Ancient solutions in silicon spread
[Verse 5]
Sparse coding principles from cortex inspire
Neural efficiency sets networks on fire
Skip connections echo brain's parallel streams
Residual pathways fulfilling visual dreams
Feedback loops and recurrence now finding their place
Biology's recursion in silicon's embrace
[Chorus]
Biology's whispers in silicon dreams
Convolution and pooling, simple and complex schemes
From edges to textures, parts to the whole
Gabor detectors emerge from the code
The brain's ancient wisdom in networks reborn
Teaching machines how vision is worn
[Outro]
From Fukushima's vision to transformers today
Biology guides us along neural pathways
Nature's algorithms refined through time
Inspire the networks in silicon's chime
55. Where artificial and biological vision diverge (and why it matters)
[Verse 1]
The camera clicks and captures light, but that's where likeness ends
CNN flows straight ahead while brain circuits bend
Feedforward stacks march single file through layers clean and neat
But biological vision loops back, makes the cycle complete
[Chorus]
Where silicon meets flesh, the pathways split in two
Feedforward versus recurrent loops, spikes versus values smooth
Backprop trains the artificial mind with gradients flowing back
While Hebbian rules and local codes keep biology on track
[Verse 2]
In primate IT cortex, some images take their time
The settling loop keeps working when first pass can't define
Those tricky visual puzzles that make the monkeys pause
Are exactly where the CNNs break their feedforward laws
[Chorus]
Where silicon meets flesh, the pathways split in two
Feedforward versus recurrent loops, spikes versus values smooth
Backprop trains the artificial mind with gradients flowing back
While Hebbian rules and local codes keep biology on track
[Verse 3]
Artificial neurons whisper numbers soft and continuous
But real neurons fire in bursts, timing synchronous
Rate codes tell us how much, but spikes encode the when
Latency and synchrony dance in the temporal den
[Bridge]
Add some recurrence to the net, CORnet-R shows the way
Closing gaps between machine and brain processing day by day
But adversarial attacks still fool the pixel-reading eye
While human vision sees the shape that textures can't deny
[Verse 4]
The global gradient flows backward through each weighted connection
But neurons can't send error signals in reverse direction
STDP and BCM rules, three factors gate the learning
Local plasticity solves the puzzle that researchers are discerning
[Verse 5]
Object recognition scales through the ventral visual stream
V1 detects the edges, IT holds the semantic dream
But cortical columns organize in ways we've yet to code
While artificial layers stack like freight along a road
[Chorus]
Where silicon meets flesh, the pathways split in two
Feedforward versus recurrent loops, spikes versus values smooth
Backprop trains the artificial mind with gradients flowing back
While Hebbian rules and local codes keep biology on track
[Outro]
Fragile networks stumble on perturbations microscopic
While robust biological vision stays philosophic
The divergence shows us clearly what the feedforward stack is missing
Recurrent loops and spiking codes are what we should be wishing
[Chorus]
Where silicon meets flesh, the pathways split in two
Feedforward versus recurrent loops, spikes versus values smooth
Backprop trains the artificial mind with gradients flowing back
While Hebbian rules and local codes keep biology on track
56. The benchmark that scores the correspondence
[Verse 1]
In the lab where neurons fire and dance
Scientists built a scoring chance
Brain-Score measures what computers see
Against the circuits, you and me
Layer by layer, V1 through four
IT cortex keeping neural score
Schrimpf and DiCarlo found the way
To test if silicon thinks like clay
[Chorus]
Brain-Score, Brain-Score, counting every match
Model to biology, see what neurons catch
V1, V2, V4, IT in the stack
Brain-Score tells us which ones bring the magic back
Correspondence measured, number on the board
Higher scores mean artificial minds accord
[Verse 2]
Feed the same image to both sides
Primate vision and computer guides
Record the spikes from living cells
Compare with what the model tells
Hierarchical features climbing high
Convolution patterns multiply
Each synthetic layer must align
With nature's own visual design
[Chorus]
Brain-Score, Brain-Score, counting every match
Model to biology, see what neurons catch
V1, V2, V4, IT in the stack
Brain-Score tells us which ones bring the magic back
Correspondence measured, number on the board
Higher scores mean artificial minds accord
[Bridge]
Normalization keeps the signal clean
Recurrent loops enhance the scene
The models scoring at the peak
Mirror biology's technique
Load-bearing motifs rise above
The architectures neurons love
[Verse 3]
When the benchmark gives its final grade
Evidence of progress can be weighed
Artificial stacks that score the best
Have passed biology's toughest test
Module ten, method ten precise
Formal proof rolled like loaded dice
The cleanest measure we possess
Of visual processing success
[Verse 4]
Neural noise and trial-to-trial change
Brain-Score handles variance range
Bootstrapping confidence intervals tight
Statistical testing gets it right
Cross-validation splits the data clean
Best performing models rise supreme
Temporal dynamics, static views
Every correlation gives us clues
[Chorus]
Brain-Score, Brain-Score, counting every match
Model to biology, see what neurons catch
V1, V2, V4, IT in the stack
Brain-Score tells us which ones bring the magic back
Correspondence measured, number on the board
Higher scores mean artificial minds accord
[Chorus]
Brain-Score, Brain-Score, counting every match
Model to biology, see what neurons catch
V1, V2, V4, IT in the stack
Brain-Score tells us which ones bring the magic back
Correspondence measured, number on the board
Higher scores mean artificial minds accord
[Outro]
From pixels raw to recognition
Brain-Score judges the rendition
Benchmark singing nature's tune
Artificial dawn meets neural noon
57. The practical takeaway for building biologically-grounded vision
[Verse 1]
Start with the retina, nature's first stage
ON and OFF channels dance on the page
When light hits bright, the ON cells sing
When darkness falls, OFF takes the ring
Color opponents red-green and blue
Yellow's the partner that makes the crew
Decorrelating signals clean
Sharpest foundation you've ever seen
[Chorus]
Build it biological, layer by layer
Feedforward's the guess but feedback's the prayer
Oriented filters then pool what you find
Receptive fields growing, position-blind
Divisive normalization keeps the peace
Recurrent loops where robustness finds release
The sweep is fast but the loop earns its keep
That's where your vision learns to see deep
[Verse 2]
First cortical stage needs oriented banks
Simple cells firing in ordered ranks
Edge detectors tuned to every degree
Horizontal, vertical, diagonally free
Pool the responses, complex cells emerge
Invariance grows as the pathways converge
Hierarchy climbing with each neural floor
Bigger receptive windows than the stage before
[Chorus]
Build it biological, layer by layer
Feedforward's the guess but feedback's the prayer
Oriented filters then pool what you find
Receptive fields growing, position-blind
Divisive normalization keeps the peace
Recurrent loops where robustness finds release
The sweep is fast but the loop earns its keep
That's where your vision learns to see deep
[Bridge]
Context surrounds every computation
Divisive control across the nation
Of neurons firing, gain modulation
Predictions flowing down like revelation
Attention carves the signal bright
While suppression dims the surrounding sight
Most models skip this backward stream
But recurrence rules the visual dream
[Verse 3]
Every level needs its normalization
Divisive balance across the population
Winner-take-all when the signal's hot
Soft competition for the neural plot
Position sensitivity fades away
As you climb higher in the processing bay
Translation tolerance starts to grow
From precise pixels to concepts that flow
[Verse 4]
Dorsal stream tracks the where and how
Motion and space in the here and now
While ventral pathway decodes what you see
Objects and faces, identity free
Two streams diverging from V1's core
Each with its purpose, each wanting more
Integration happens where pathways meet
Making perception whole and complete
[Chorus]
Build it biological, layer by layer
Feedforward's the guess but feedback's the prayer
Oriented filters then pool what you find
Receptive fields growing, position-blind
Divisive normalization keeps the peace
Recurrent loops where robustness finds release
The sweep is fast but the loop earns its keep
That's where your vision learns to see deep
[Outro]
Task performance ain't the only game
Faithful modeling stakes its claim
Structure matters, biology's wise
Recurrent feedback opens the eyes
Back to Home