Visual Function and Dysfunction Flashcards
What is the evidence for Face cortical patches?
- Stimulation of these areas during surgery causes face visualisation when no face present (e.g. on a basketball)
- Damage to these areas causes inability to recognise faces (prosopagnosia)
Are observations from lesions in human patients useful for understanding?
Yes: provide causal information on the function of the area (via negativa)
* Are many different cases across the brain
* In humans! (realistic model unlike animal models) and can perform human tests of function.
No: lesions in humans are not controlled
* Very low spatial resolution (not discrete)
* Difficult to observe the damage a lesion has caused (ablation to whole area? Reduced activity?)
* Permanent
* No repeats
Evidence for object sensation in lateral occipital cortex
Electrophysiology studies in live animals being shown different visual images.
Simple shapes detected in occipital: e.g. lines of light
Complex requires convergence of many sensing cells to build image: e.g. face detection cells in IT are selectively responsive to particular faces
How was bi-stable perception hypothesised?
Bi-stable perception is the phenomonen of switching between two distinct visual images when an ambiguous stimulus is presented. These images are mutually exclusive and not mixed but perceived separately.
Tested:
* Macaques trained to press lever A in response to image A and lever B in response to image B
* Present a macaque with different image to each eye
* Record which lever they press (Sheinburg et al 1997)
((Suggests perception occurs after image processing)).
What is the evidence from bi-stable perception experiments for that IT neurons are involved in visual perception?
- Electropysiology recordings show activity correlated to specific ccomplex images
- Activity recorded just before visual perception reported (i.e. cells responsive to the face image fire when face reported) (90% predictivity)
Evidence linking IT neurons to perception (list experiments)
- Damage (lesion) to IT can leave sensation (image processing) without perception
- Electrophysiological recordings show high correlation between cells responsive to particular images and reported perception (Schienburg bistable perception experiments)
- PET scans (using 18FDG) highlight a lateral temporoccipital region active when presented with objects (famiiliar or novel) rather than scrambled lines
Describe an experiment to observe the difference in brain activity between seeing vs. to perceive
Distinguish between feature extraction and perception:
* Present drawings of familiar objects (e.g. key, flower etc…)
* Of novel objects (perceivable as an object but not recognisable as known)
* Of scrambled lines (not perceivable)
Observe activity (e.g. via PET)
What is the evidence for face-specific vs. object specific areas in the IT?
Non-human primates:
* Electrophysiological recordings showing cells specific to faces or objects
* Bao et al: one step further, used this to characterise organisation of object recognition, finding specialised regions (including face)
Humans:
* fMRI (BOLD) studies show facial regions highly selective for face presentation as opposed to object (BOLD signal much greater)
* Object-preferred or face preferred areas (since not completely selective). FPAs more selective than OPAs.
What is the evidence that object and face perception are achieved separately?
- Different physical locations (fMRI studies, electrophysiology studies showing face and object preferred areas)
- Agnosias to specific groups of objects (e.g. prosopagnosia, inanimate objects only…)
Theories on the organisation of the anterior IT
Idea of FFA well characterised
Bao et al: even IT ‘no-man’s land is organised
* Systematic electrophysiology recording in macaque brains with presentation of hundreds of types of object
* Found organisation on physical features: animate/inanimate, stubby/spiky
Might explain agnosias o specific catagories (e.g. just inanimate objects)
What are the criteria for an object memory neuron (i.e. neuron which represents a physical object)
- Responds selectively to seeing the object
- Is active when object is recalled (in absence of visual stimulus)
- Response is acquired by selective learning
What is the evidence for pair-coding neurons?
Naya et al 2003 (electrophysiology recordings)
* Objects were associated with each other in pairs
* Some neurons in A36 showed maximal response to one object and a less (but significant) response to its object pair. Quantified using a ‘paired coding index’ (response to pair/response to other pairs)
* This is selective to only the paired object
* Contrast to single coding neurons which respond selectively to one object
* More pair coding neurons in A36 compared to IT
* Coding index increases with learning (learning dependent development)
* Recalling same object also stimulates activity
= matches critera for a memory neuron!
What are single coding neurons?
Selectively activated for a specific object.
(shown throuh electrophysiology)
How do single and pair coding neurons interact? How was this investigated?
Single coding neurons (in IT) activate pair coding neurons (in A36).
Explore using paired electrophysiology recording:
* Identify cross-correlating units
* Spike seen in SCN prior to PCN ‘target’
What is the difference between single/pair coding and recall neurons?
Coding neurons are active during the cue period (=’cue holding’)
Recall neurons are active during the delay period
A neuron can be either (cue holding always relays to recall neuron) or may be both.
What is some experimental evidence showing bidirectional signalling on recall neurons?
Two visual images are associated (through learning):
* Viewing one activates specific single recall neurons
* These activate the learned pair coding neuron
* The pair coding neuron then activates the single coding neuron of the associated image
* Suggesting an individual single coding neuron can be activated by bottom up stimulation from V1 or top down ‘recall’ (mediated through A36)
How was the progression from view-selective to view-invariant coding cells investigated?
Freiwald and Tsao: experiments on face cortical patches of macaques
* Identified several temporal areas highly selective to facial images
* Areas: PL, ML and AL and AM
* Took electrophysiology recordings from these areas
* ML cells respond only to faces at a particular angle (e.g. straight on)
* AL patch show mirror-symetrically view angle invariant activity (e.g. respond to side on profile either way)
* AM are view angle invariant (respond to a face at any angle)
Activity of these cells significantly higher for known objects (learned)
What is the experimental evidence for top down activation of recall neurons?
(Causal) Macaque split brain surgery (Tomita et al):
* Repeated association tests with both objects presented to one eye (one side of the brain (i.e. association coded in hemisphere))
* Association known when object presented ot other eye
* Achieveing this requires relay through frontal areas (top down linking)
* Latency is longer e.g. 100ms (more connections to pass through) contralateral TE → A36 → OFC → ipsilateral A36 for recall
(Correlational) MEG human studies:
* Activity in OFC detected in human trials of visual associative learning before visual recognition cortex activity.
How might the frontal lobe have top-down modulation? (Give model)
Bar et al model hypothesised:
* OFC receives low quality image and suggests ‘possible objects’ to visual cortex
* Where these guesses and the visual image match, there is accurate recognition.
Is there heterogeneity amongst pair coding neurons?
YES: Naya et al 2003 (as part of pair-neuron discovery):
* Evidence for 2 sub-groups in their 25 identified pair coding neurons using cluster analysis
* Both have trend of a transient response which declines to a stead level
* Type 1: faster but less selective - show same response time to object and pair cue and small response to ‘other’ objects (i.e. not pair) (lower PCI)
* Type 2: slower but more selective - have longer response time to paired stimulus but no response to other objects (higher PCI)
What is the evidence for selective convergence model of pair associative memory?
Idea that visual memory arises from addition of multiple sensory and visual inputs.
E.g. a pair coding A36 neuron only becomes active when both a red circle and white circle are presented together (not separately).
Supported by:
* TE activity first then delay followed by A36 activity (TE driving A36)
* Large increase in pair coding neurons in A36 (Naya)
What are the theories about how pair-coding neurons arise?
Selective convergence theory: single codign neuron of object 1 + 2 → pair coding neuron
Direct driven model: neuron coding object 1 → pair coding neuron → object 2 neuron
What is the evidence for hte direct driven model of associative pair coding?
Suggests that: neuron coding object 1 → pair coding neuron → object 2 neuron. Therefore presentation of one object induces visualisation of other.
Evidence:
* There are some pair coding neurons in TE
* Perirhinal lesion experiments eliminate pair coding neurons in TE (require some A36 to TE feedback)
Repitition of a visual association learning task format increases learning speed. How does this work?
Participants become faster to reach same accuracy:
* Could be a memory schema - repetition builds a framework which new information can be overlaid onto (Tse et al 2006)
* No causal/mechanistic explanation is agreed on
* Or could be the upregulaiton of ‘learning genes’ in the visual cortex, pre-disposing for faster associative learning.
What are the basic visual memory functions?
Familiarity (‘I have seen this before’)
Recognition (‘I know what this is’)
What are memory engram cells?
Particular group of cells in the perirhinal cortex which seen to be causally involved in the coding and representation of a visual object with their associations.
Sufficiency: activation causes memory expression independent of stimulus
Necessity: selective ablation decreases memory expression
Evidence for the sufficiency of memory engram cells for encoding visual associative memory:
Selective activation causes increased memory expression (Kim et al 2018):
1. Learn association between visual image and shock (fear respnse seen)
2. Uses tetracycline inducible conditional expression system (Tet-OFF)
3. When doxycycline is not administered, ChR2-EYFP is expressed
4. Removing doxycycline during training window means only active cells express ChR2-EYFP
5. Allows selective control of cells active during learning
6. Optogenetic activation of ME cells (in perirhinal) increased accuracy of context recognition and memory retrival (expression of fear memory)
Evidence for the necessity of memory engram cells for encoding visual associative memory:
Selective ablation causes decreased memory expression:
1. Object-object association task learned (one object associated with shock so induces fear response)
2. Memory engram cells identified by high CREB levels (strongly associated molecular marker for learning)
3. Cre used to ablate high CREB cells with DTX causes decrease in memory expression (fear response)
4. Compared to randomised ablation of same cell proportion
Present evidence that perirhinal cell activation mediates visual recall in macaques
Tamura et al 2017:
* ChR2 and fluorescence promotor delivered to perirhinal cells (AAV)
* Activation and fluorescence distribution confirmed later using histological analysis
* Monkeys presented a familiar image which slowly becomes pixelated until unrecognisable then unpixelated to reveal new image.
* Familiar object chosed with increased probability at same noise level (as new object)
* Optogenetic activation caused familar object selection at higher frequency even if new object presented. (i.e. they are recalling familiar object)
What pieces of evidence are there for perirhinal cells being involved in visual memory?
Rodents:
* ME cells are sufficient and necessary for memory expression
Macaques:
* Perirhinal cell activation mediates visual recall (Tamura et al)
What are the problems translating from rodents to humans for visual processing experimentation?
Rodents have different anatomical structure:
* Smaller visual cortex
* 28% cortex as opposed to 80%
* No developed frontal lobe (known to be invovled in higher level processing)
Rodents rely less on vision:
* Much lower visual acuity
* Not used as primary sense
Experimental considerations:
* Human paradigms hard to replicate
* E.g. difficult to teach mouse many different object associations, therefore fear conditioning often used
What are the major molecular changes associated with learning?
Learning related genes upregulated:
* Brain derived neurotrophic GF BDNF - expression increases in perirhinal cortex after association learning
* Cyclic AMP response element BP (CREB) - necessary for memory formation (Han et al)/Johansen ‘CREB ablation diminishes learning ability’
Evidence for the role of OFC in visual identifcation and memory
OFC important for object memory (in conjunction with perirhinal) but not perceptual identificaiton:
* DMTS task with short (0.3s and long (3s) delay
* Inhibitory DREADD injected into OFC
* Caused significant impairment to DMS task (no effect for 0.3s but significant for longer delay)
* OFC inactivation increases anterior ventral temporal cortex activity - normally suppressing?
Major limitation:
* CNO activation was of OFC cells not OFC to aVTC pathway. Therefore damage could be disrupting another pathway.
* Should express DREADD in connection between the two
