spatial Flashcards
HM describe what happened to him
- don had intractible epilepsy
- surgeon Scoville removed about 7-8cm of the medial temporal lobe
- extracted all of that material by suction
- damaged amygdala, erihinal cortex, postrhinal cortex but primarily taking out the hippocampus
Scoville and Milner paper (1950’s)
what did they argue
- the more the hippocampus was damaged, the more amnesic the patient
- concluded the hippocampus is the cause of amnesia
- thus the hippocampus MUST be a memory structure because if you remove it you get amnesia
important to note the extent of the lesion was wrong
if we remove the hippocampus what type of memory is fucked?
episodic memory in particular
HM couldn’t form new autobiographical memory after the event
lived the rest of his life as a dense amnesic
if the hippocampus is involved with (episodic) memory what happens if you put in electrodes and recode from them?
shockingly, discovered place cells.
different cells code different parts of the environment
cells form a cognitive map of the environment
why is it so shocking that electrodes recorded in the hippocampus produced place cells
because we thought the hippocampus was a memory structure? now its a spatial structure? argggh whats the truth?
how do place cells relate to episodic memory?
how do place cells help us form episodic memory?
place cells in rat hippocampus vs episodic memory of human hippocampus - how can we explain the differences?
Just a different structure in rats vs humans ?
nope!
- much of the neuroanatomy from rats, monkeys and humans have preserved the regions - H structure just flipped 180 degrees with evolution
- place cells are also found in monkeys (Ludvig et al., 2004) and humans (Ekstrom et al., 2003; Miller et al., 2003)
evidence suggests that navigational and episodic functions are consistent across spiecies
limitation of patient studies in humans - electrodes recording the hippocampal place cells
your recording form a dysfunctional area (its about to be taken out)
place cells in rat hippocampus vs episodic memory of human hippocampus - how can we explain the differences?
Hippocampus is not a single structure - one part does episodic and one part does spatial memory
argue for this case
Well the hippocampus does contain many different structures. See changes in anaotmical structure itself, but also changes in anatomical connectivity
- e.g., DG - receives lots of input then only outputs to CA3. this is the input structure to the hippocampus
- subiculum hippocampal output - receives a lot of information from CA1 with modest output.
- basically, everything the hippocampus backs out to the cortex is via the subiculum
Thus maybe different types of input affect the function of a sub-region…
Describe the trisynaptic pathway
place cells in rat hippocampus vs episodic memory of human hippocampus - how can we explain the differences?
Hippocampus is not a single structure - one part does episodic and one part does spatial memory
argue against this point
there is no strict divide between one structure doing place cells and another structure doing episodic memory. we see place cells throughout.
also, we don’t see a lesion affecting episodic processes vs another lesion inducing episodic impairment
- patient RB - through loss of oxygen in surgery - lost only the cells of CA1 (bc those cells are very dense in glutamate receptors, these are highly energy requiring, if you remove oxygen those cells die before anything else)
- but this guy still had amnesia - so can damage small part of H and still get amnesia
- althought here does appear to be anatomical division, doesnt explain the place cell vs episodic memory debate
Why should we strive away from episodic and semantic labels on things?
as a cognitive neuropsychologist, your aim is to put cognitive function onto biological units. Try to break the world down into cognitive functions. and this cognitive function should map onto a biological structure.
That mapping should reflect the computation of that region
so in doing this computation you achive these cognitive ability
terms like episodic and semantic memory are very arbitrary concepts - difficult to map these onto a structure. were too fixated. on things like episodic and semantic memory as terms which are very difficult to map onto biological processes
generic terms bad - reflect multiple component computations. yet were trying to define them ono a single region of brain
reconciling episodic and spatial navigation in the hippocmapus
Relational processing and memory space
what do both episodic and semantic memory have in common?
- all objects are related to something - this glass is on the table
- there is a relationship with time, “yesterday” or “just after we went to the beach
- in building relationships between items and spaces in time we form an event memory
similarly with navigation
- walk from part b along this line to point B then along this line to point C
- allows you to learn a geometric relationship between those things
- can say this is starting point and this is the end - can only work that out because you’ve formed a relationship between the different points
Eichenbaum argues that both navigation and episodic memory share this feature - you try to relate things to one another. need to bring the relationship between things together to achieve either of these goals
eichenbaum quote reconciling link between episodic memory and spatial functions in the hippocmpaus
rather than thinking about navigation in a spatial world think about how your brain is full of memories. the hippocampus is helping you relate between things you can navigate your memories
if asked what you had for breakfast this morning vs what did you have for breakfast yesterday morning you can answer both of those by forming relationships between the information you need to know: breakfast + this morning; breakfast + yesterday morning ( a different relationship)
so basically you navigate memory space rather than navigating geometric space
evidence support for relational processing?
Wood et al (2000)
- found not all place cells just represent space
- not just a map of where you are
- it can also be a map of where you are in relation to what your doing
deign
- figure of 8
- rat told travel up line go left, then travel up and go right - do this repeatedly
found
- place cells exist that represented all areas of the maze
- cells representing the middle of the maze - which n travel down in both circumstances
- found that some cells (blue ones) oNLY fire when your in that space and about to turn left
- other ones (red ones) only fire when you are in this space and about to turn right
no longer cells that represent only spatial information. now represents information about where you are and its relationship with what you’re going to do
how did wood et al (2000) study support the relational processing theory?
because if these cells are actually not interested in where. you are but the relationship between things
and space comes out because its easy to see
but actually if it’s interested in other types of relations as well i.e., what you are about to do then we can understand the hippocampus as a structure relating things together
does the relational processing theory (Eichenbaum) also account for the issue of time?
Eichenbaum (2013)
- argued he was showing hippocampal place cells also represent time
design
- animal ran around track
- here animal sees object, dalay period, then knows it will get an odor
results
- during the delay
- when the animal is thinking oh i’m going to get odor 1, during the 10-second delay, we see a cell fires at the first second then switches off, then another cell fires the second seecond then switches off and so on
so he argues just as place cells mark a space in an environment, we have time cells that mark different time in environment
when it comes to episodic memory - most people would argue that time is a critical part of episodic memory. why?
- firstly bc you can locate a memory in time - e.g., what you have for breakfast today vs yday vs last sunday
- secondly, when you remember an event - most of us have an internal visual image of the event as you experienced it, you know you arent hallucinating. you know its a past memory.
- what edward tolling called “a sense of pastness”
did Eichenbaums odor study acc find time cells?
yes
- so you might argue oh its just computing space - but he did another study where the animal was on a treadmill and you still see these cells fire so not doing space but time
yes
- argues against precession because in that you can vary the length of the delay
- find some cells at the start are fairly consistent so regardless of whether its a short or long delay the same start cells switch on and off
- however, cells near the end of the delay have a much longer firing
- the shorter the firing range of a cell the more discrimintory your abiltiy is - fewer cells on at a time, so you know when these few cells are active were at this half a second period
- but near the end - last 4 seconds lot of cells are active. difficult to determine whether youre at second 4 or second 8
- no matter how many times you do this study your disciminatory ability is better early one than later on
- argues phase perception isn’t what’s going on because you would expect that firing to be fairly standard throughout the delay
- you wouldn’t expect it to be delay-dependent - shorter at the start and longer firing at the end
yes
- in his study 2014
- found if you extend the delay it alters the behaviour of the start and end firing rates
- so instead of thinking in absolute time - 1 second 2 second, 3 seconds etc thikn of it in relational terms - shortly after the onset, shortly before the offset - makes more sense
- those are relational terms of time
- purple lines - its expecting this to end
so
- “time” - many different types of time
- problem when you call something a time cell, what sort of time is it actually measuring?
- all of this makes more sense in a relational time manner
- okay its started now, now im still going on this treadmil, okay still going getting a bit bored, okay now its nearing the end woo
what do we need to consider when using IEG’s to look at
- well there’s always gonna be bare cells firing - so your control group needs to be VERY good. like with fMRI
- need to ensure you are comparing it with something comparable so you aren’t making mistakes in the reduction
- also need time for these genes to express themselves - so you need the animal to do something then wait a bit
- then you need to kill the animal - as you cannot image this through the skull - get slices and inspect them
- gives you a kind of spatial detail that yuo dont get with fMRI as were looking at individual cells that were active during tha time
so its a powerful technique with some limitatinos
immediate early genes (IEG) what are these?
- when cells fire, a cascade of proteins that occur
- genes switch on - produce proteins in order to carry out the function of that cell when it fires (to return, to recycle proteins, to release NT’s)
- there are immediate early genes that switch on as soon as cell fires
- basically, genes that activate rly quickly to say right. this cell is now active, and starts those protein cascades
- if you attach a dye to those genes when they switch on, means you get a snapshot in that cell.
- so could look at a whole slice of brain, the cells that are dyed mark the genes are active
- and the genes are active in the cells that have just fired
Wan et al (1999) - 1st part when shown the novel and familiar
study design
- paired imaging
- rat’s eyes are on the side of their head
- rat puts nose in hole and each eye is exposed to a different screen
- one hemisphere processes one image e.g., familiar picture and the other processes another e.g., novel picture
- compare activation in the same structure in each hemisphere
- subtract activation from one hemisphere from activation in another hemisphere - the difference you are looking at here is the difference between novelty and familiarity
- very clever design, animal can act as its own control
results
- activity in the hippocampus does not distinguish between novel and familiar objects
what is postrhinal in the rat called in monkey/humans?
parahippocampal gyrus
what happens in the hippocampus when you look at a picture?
fuck all - activity in hippocampus is same when you look at novel vs familiar pictures
other areas however e.g., perirhinal cortex care about seeing novel pics vs familiar
adds weight to the Angleton and brown perspective that semantic memory relies on structures outside of the hippocampus. it’s evidence like this. that object memory recognition isn’t hippocampal - occurs in areas like. theperihinal cortex
Wan et al (1999) - 2nd part when shown items in novel and familiar locations in space
- now what’s novel or familiar is not just the picture - it’s the relationship between picture and location
- NOW the hippocampus cares
- 2 subregions of h - DG and subiculum - activate MORE with familiar arrangement of items compared to novel arrangment of items
- CA1 -opposite - likes novel arrangements more than familiar
so same structures that previously didn’t care about the objects, care about the same object’s relational positions in space. strong evidence support for Eichenbaums relational processing claim. shows H doesn’t care about things, but relationships between things
Gaffan’s 1994 - scene memory hypothesis
Argued that when you remember an event you remember the whole scene. scenes allow you to discriminate one event from the other.
Discriminating breakfast from today vs yesterday is not done through time alone, but the fact that the WHOLE SCENE is different - different people, different things happening etc.
how did Gaffan test scene memory hypothesis?
Gaffan (1994)
- presented monkey touch screen
- two foreground objects “X” and “O” in very specific locations
- background of different colours/objects
- animal learns that one of these two objects if they touch it they get a food reward
- each of the two stimuli (each problem) is against a consistent scene, so X in THIS SCENE against THIS BACKGROUND will always be in top left
Gaffan believed that would allow the animals to remember these problems… and he was right
By the 2nd trial, animal got down to 20% error. and almost all of those errors were from things previously got wrong - animal made similar error. never made error when it was something they got right. by trial 8 n perfect.
Gaffman scene study - how did he show performance was hippocampal-dependent?
- Lesioned the fornix - disrupt hippocampal functioning
- significantly slowed down how fast the animal learned this skill
So putting objects in specific locations in specific scenes allows for rapid learning and that rapid learning is hippocampal-dependent
Which animals display scene memory dependent on the hippocampus?
- monkeys
- rats
- humans
also shown dependent on the hippocampus as
- amnesic n show impairments in scene memory
- animal models of AD show severe impairments in scene memory
Spontaneous object recognition (where-what-which task)
how does a hippocmpal lesion affect this?
fornix lesions showed this task was hippocmpaal dependent.
- 0 is chance and 1 is perfect peformance
- animals with no lesion (sham) = good at task
- lesion of perihinal cortex (pe) = good at task
- lesion of postrinal cortex (po) = good at task
- disruption of hippocmapus through = CANT do task
the what where which task. -could just argue that the reason the hippocampus is invlved is bc H does spatial and theres a spatial element to this task.
argue against this
- instead of making rat do what where which
- they do what where task
- completely fine doing this despite hippocampal lesion
Maguire and Mullay quote on scene memory
any evidence to support this?
how can we test SCT in humans? understand amnesia using SCT
- so we know that episodic memory and spatial navigation depend on the hippocmapus
- if SCT is true then amnesic n should be impaired in any skill htta requires sccene construction
- so we can test other things that require scene construction and see if these skills are also impaired in amnesic n
how is the scene in the minds eye of an amnesic patient
somewhat disturbed - “squished”
can amnesic n imagine scenes?
no
Hassabis et al., (2007)
spatial coherence of the imagined scene in amnesic n
how do we test performance objectively
spatial coherence index - (this only part of it - full experiential index captures full range of information)
- ask n to judge how spatially coherent the scene feels
- can you move around it? hows the geometry is it squashed/disturbed in controls - very spatially coherent image
- however for amneiscs - get a negative coherent index
- means they are describing things that aren’t spatially feasible
- however… 2 amnesics show good spatial coherence.
- interesting…?
patient 1 and Jon really good at the scene construction task - scored highly on the spatial coherence index
doesnt this disprove the SCT theory?
we can explain these two outliers
- both had only half their hippocampus damaged
- so, while half a hippocampus isn’t enough to stop n being amnesic is it enough to have them spatially construct scenes
Mullally et al., 2014
- neuroimaging while they imagined hypothetical events
- told to imagien things, tone sound told them to open eyes and stop
- then they rate how difficult it was
PO1 and jon results
- the remaining bit LEFT and RIGHT hippocampus was active during the task
- so whats left is able to support some construction of scenes
jon
- remaining LEFT hippocampus was being used
- also activated the retrosplenial cortex more than P01 and controls
- regions that we wouldn’t normally use
- is this compensatory activation for the lack of R hippocampal activation?
is SCT right?
well SCT would predict that amnesics should struggle with other, scene-based tasks, and that is what we find true so… can add this to other evidence support for the theory
what other ways have people investigated scene construction in patients?
boundary extension effects
- asked to draw image
- controls draw it zoomed out 60% of its true size
- amnesic patients consistently draw an accurate depiction
the boundary extension effect - where amnesic have better memory than controls - is it true if n just see pictures
yes! Mullaly et al., 2012
- n see picture
- then see second picture asked if it’s a) same b) closer up or c) further away
- controls confidently (and inaccurately) say its closer more often while experimental patients always say
extension how can we explain the boundary effect
- when viewing close up scenes we automatically extrapolate beyond the physical edges of a scene
- e.g., with a cat, understand the tail extends and behind that tail extension theres more grass prob beyind
- hippocapmus understands and does the computations for this
- so you construct boundaries using your understanding of how the world exists
- v cool - the BE effect (the extended scene) is what you acc remember. Compare memory against the picture you constructed, not the one you seen
its the extrapolating additional details beyond the boundaries of the scene that patients might not do
Chadwick et al., 2013 - what did this study show
the boundary extension effect is hippocampal dependent
- if you are about to make a BE error then the hippocampus is active at the time of ENCODING
- not happening in the reconstruction of the memory - but at encoding
- when first seeing it - you immediately construct the zoomed out image in your head
