Animal Cognition: Comparative models of working memory

What is cognition?

Core abilities / executive functions

Scala Naturae “Ladder of Being”

Fish can demonstrate working memory even though 'low intelligence'

Frogs have demonstrated cognitive flexibility such as learning rules and applying them

Lizards can display inhibitory control

Is insight-related cognition solely a feature of mammals? (test of the scala naturae argument, next lecture)

Birds and mammals are evolutionary quite distant  (300 million years!)

What’s changed? Neural Structures? Cognitive Abilities?

We have seen brief examples of core cognitive functions across a number of species, lets explore this!

Cortex is special or is it?

…lamination defines a new computationally advantageous module (Montagnini & Treves, Brain Res. Bull., 2003)

… differentiation of the neocortex … opened the way to a spectacular development of cognitive … skills (DeFelipe, Front Neuroanat, 2011)

Cortex has 6 layers of difference cells which overlap to allow for quick communication

• Something we have which animals do not have

• Make us more intelligent right? Not necessarily

Neocortex - lamination

Prefrontal cortex is especially special

Prefrontal lobe associated with core cognition and executive function (Yuan & Raz, 2014)

Damage to Prefrontal Regions associated with deficits in executive function (Løvstad et al., 2012)

• phinneas gage

Open questions – dissociating between executive functions, localisation of lesions in patient studies.

Comparative approach - birds vs humans

-Birds and mammals are quite distant from each other in evolutionary terms (“scala naturae” argument)

-Birds do not have a neocortex (“Cortex is special” argument)

Working memory

Action potentials - spike trains

Main Form of communication between neurons

Electrochemical Signal

Brief, Stereotypical changes in the membrane potential of a neuron

Due to opening and closing of ion channels

Primarily due to flux of sodium and potassium

Governed by gated ion channels

Delay activity - a neural basis for working memory?

Delayed response task for monkeys (Niki, 1974)

One of two cue lights is illuminated for 1s (left or right)

Then the cue illumination is turned off for 2-3s (delay period)

Then the monkey has to choose between left and right

Delay activity in example units in the prefrontal cortex

Delay period is crucial - shows us if the trained monkey is holding onto the info

• Correct more than chance would estimate

Should be a lot of activity in the PFC

Monkey must remember the light that turned on

When the cue comes in there is a spike

There is then an increase in neuronal activity during the delay of the light being off

• This is sustained to the choice

Reflective of the monkeys retaining this info

Choice - rapid drop off of neuronal activity after he choice is made as the info no longer has to be retained

Initial spike train, sustained spike train and then a drop off

Dopamine and classical conditioning

R = reward - orange juice

• Increase in dopamine

After consistent pairing of CS and US - response occurs at CS not R

• Reacting to anticipation of reward

Dopamine is a neuromodulator

Respond to sensory stimuli which predict future rewards

Any study will need to take into consideration rewards involved in the training of the task

Initial conclusions

No complicated things happening in the study which could be argued to just be STM not WM

Is cortex special - the avian brain

More similar to the makeup of our subcortical areas

The architecture of most of the brain is similar to the basal ganglia

• Instinctual responses

This would lead us to believe they think in a more instinctual way

Green - higher cognitive functioning areas

This has been reconsidered and redeveloped over the years

• The thought that different architectures of the brain supports different types of the brain

The thought that primitive sort of structures lead to primitive behaviours

• This has been reconsidered

The pallial layer may be supporting some of the higher cognitive functioning

It is however different from the cortical structure we have

Working memory in crows

Working memory In crows

• Delayed match-to-sample task (MTS)

Disentangle working memory from motor preparation - location of stimulus changes randomly

Chance for reward was equal for all match items, randomised, reward expectation negated

Nidopallium Caudolaterale (NCL)

“Pallium” is a neuroanatomical term for the grey and white matter covering the cerebrum

Nidopallium = nested pallium

Prior to 2002 the nidopallium was referred to as neostriatum, indicating a more primitive function

The nidopallium consists of many subregions, one of them is the caudolateral part, the NCLSee this in most bird species

 

The Nidopallium may be analogous to the prefrontal cortex of the human brain in terms of its function

• In terms of structure, anatomically in a different  location but may be performing the same tasks 

  See this in most bird species

   

  The Nidopallium may be analogous to the prefrontal cortex of the human brain in terms of its function

  • In terms of structure, anatomically in a different  location but may be performing the same tasks

Caudolateral – “of the side and tail”

See this in most bird species

The Nidopallium may be analogous to the prefrontal cortex of the human brain in terms of its function

• In terms of structure, anatomically in a different  location but may be performing the same tasks

Is the avian NCL analgue to the mammal PFC

Analogue – same or similar function even if structure is different

Working memory task in homing pigeons

Directed Forgetting procedure

Single Unit recording in NCL during task performance

Black lines here indicate electrode tracks

Numbers refer to anatomical coordinates of coronal sections

Examined honing pigeons and pigeon behaviours

Examining whether pigeons could engage in WM tasks and if there was a neuronal response in the nidopallium caudolaterale (NCL)

• Whether it provides or support that cognitive function

Similar to the research on monkeys but in birds

Absence of lamination - tells us that this might not be necessary

Inserting an electrode into the brain during tasks which require WM

NCL vs PFC study in homing pigeons

ITI - inter trial interval

• Start off with a break

• Nothing going on

• Pigeon is situated within the box

Sample

• Stimuli presented

• Animal pecks cue

• Image of a circle

• The pigeon has to code that pattern

• Array of patterns which the pigeon has to encode

• It is specific - not one or the other

Cue

• Remember cue - audio cue which indicates that they need to remember the info - high note

• Forget cue - audio cue which indicate that they need to forget this info - low note

Delay

• 3s delay

• Pigeon must hold onto the info

Comparison

• Test period

• Must select sample from 2 choices

• Discriminate between 2 targets

• Figure out which matches most closely to what they had to remember

Provided a food rewards on correct selection

NCL vs PFC study in homing pigeons - careful in interpretation

Abolished delay activity could be linked with reward prediction – food reward

Neurons could change firing rate in response to forget, because no reward!

Discern between confounds of WM and Reward prediction

Neuron could be involved in both!

Remember

• Stimulus presented and there is a spike in neuronal activity

• After cue is a consistent spike train

• Sustained delay activity

• Infer the pigeon is holding onto that info and maintaining it

Forget

• Stimulus presented and there is a spike in neuronal activity

• After cue there is a return to baseline

• This is reflective of not processing that info (WM)

NCL may be a locus of info

• WM behaviour

Pigeons have WM but the architecture which makes up this part of the brain is entirely different to that of a human brain

Working memory in crows

Go stimulus

• Tells them they are about to engage in a task

Pre sample black screen

• Neutralise what they have seen before

• Masking any impression of anything previous

Presented with sample and then delay

Choice

• In every trial the crow has to make a choice

Randomised reward

• There is a chance for a reward and this is hard to predict because time between rewards varies

Crows perform task at very high level!

Very good at cognitive tasks

Example for “sample-selective” neurons

Response of neuron – active during selection period

This means that the neuron distinguishes the four different sample stimuli with its firing rate

Here: highest firing rate for stimulus S3 (with gradual differences to the other stimuli)

Delay activity is not simply in response to the task itself

The crows for each sample they demonstrate a preference in the firing rate of the neurons

Neurons fire differently in term of their rates depending on what they are observing

• 4 possible items they are observing

Selective discrimination of stimuli – non-arbitrary

Example for “delay-selective” neurons

• Delay-selective means that the response of the neuron is selective during the delay period

Note that during the sample period the neuron did not distinguish the stimuli

Here: lower activity for stimulus S4

More than half - either sample or delay selective or both

Advanced analysis on the relation between neural activity and task performance

Perform task very well overall – but what about the errors!

Bottom line: in error trials the firing rate of the neurons does not distinguish the different stimuli as well as in correct trials

This is true for both sample and delay selective neurons

Firing rate when stimulus and delay can predict whether they get the trial right

• If they have a preference and a very high firing rate they are likely to be able to identify it

Advatnages and disadvantages of these methods

The bee brain

Small (~0.5mm3)

Only ~1,000,000 neurons (humans 16,000,000,000)

But large compared to fruit flies (100,000 neurons)

Not a simple, hard-wired structure

Long tradition of honey-bee physiology, navigation and communication

More recently studies on learning and cognitive capabilities

Working memory in honey bees

Varied the length of a tunnel which the bee had to fly through to do delay

Sample which they have to encode they can move up and down the tunnel

• The further along the tunnel it is (to the entrance) the further the bee has to fly before it makes its choice and vice versa

Assess the WM capacity of the bee

Sample is what the bee has to remember

Geometric shape as it flies through the tunnel

As the bee progresses through the tunnel it then has to make a choice between lots of different stimuli when it gets to the end

Delayed match-to-sample task

Visual sample stimulus in the tunnel

In decision chamber choice between the sample stimulus and a different (incorrect) pattern

High performance levels for short delays (~75% correct for 1.24s)

Varying the distance d2 between sample and decision (= varying time interval = varying difficulty)

Performance indicates “working memory” of up to ~6.5s

• higher rate than chance

More difficult memory in honey bees

Let’s make things more difficult for the bees...

What if there is also the incorrect pattern somewhere in the tunnel?

So now either before (IP1) or after (IP2) the sample stimulus the incorrect pattern is shown in the tunnel as well

The correct sample pattern is always presented at the same distance (120cm) from the entrance

Learning test 1: Baseline condition (without incorrect patterns)

Transfer test 1: Incorrect pattern (IP1) presented in front of the correct one

Transfer test 2: Incorrect pattern (IP2) presented after the correct one

Learning test 2: continuation of training

Transfer test 3: correct pattern suddenly at the IP1 position (ie 50cm) and incorrect pattern at IP2 (170cm)

Performance in all cases above chance level, except for the Transfer test 3 control

Working memory

The representation of items held in consciousness during experiences or after retrieval of memories. Short-lasting and associated with active rehearsal or manipulation of information (Miller, 2000).

To what degree does this match the memory task in the bees?

• Important contribution of Psychology to other fields by providing clear terminology and requirements for cognitive abilities 

Potentially working memory – but more properties of STM – Grey Area

Some more conclusions

Despite the small size of their brain, also insects have a variety of cognitive abilities.

Bees might have a “working memory” of several seconds.

But it remains currently unclear to what degree the active memory content can be manipulated as e.g. in primates