Content 4

Learning Objective

  1. What is the difference between the 3 theories of working memory and their hypotheses of short term memory stores?

  2. Explain the main result of the Ranganath study and the interpretation of that result

 

Short term vs long-term memory

  • Short term: memory for info currently held "in mind"

    • Limited capacity

    • Predominately involves the frontal lobes

  • Long term: refers to info that is "stored"

    • Unlimited capacity

    • Predominately involves interactions between the temporal lobe and the cortex

 

The difference between short-term memory and working memory

  • Short term: passive retention of info temporarily held in mind

  • Working: is the system for the manipulation of info currently in short term memory

 

Working memory

  • Most working memory models assume the main storage site of info is long term memory

    • The long term memory system is not in the frontal lobes

  • The prefrontal cortex is mostly responsible for working memory

  • It keeps info active and/or manipulates the active info according to current goals

    • Limited capacity

 

Single - unit recording and spatial working memory in non-human primates

 

 

Atkinson and Shriffrin: Modal Model of Memory

  • Once info enters the brain, it must either be stored or maintained and the info which is stored goes into 3 distinct memory systems

  • Info is processed in a serial manner

  • 3 distinct memory systems:

    1. Sensory store: sensory info enters memory

    2. Short term store (working memory): info from the sensory store and long term store is received and held

    3. Long term store (long term memory): where info that has been rehearsed in short-term store is stored indefinitely

 

Baddeley and Hitch model of working memory

  • 3 short term stores (blue) that interface w/long-term memory (green)

    1. Visuospatial sketchpad

    2. Episodic buffer

    3. Phonological loop

 

  • Central executive: an attentionally limited system that selects and manipulates material in the subsystem, serving as a controller that runs the whole show

  • Episodic buffer: a storage system that can hold limited info in a multi-dimensional code to interface w/long term memory

 

Visuospatial working memory

  • Holding visually and/or spatially encoded items and arrays in minf

  • Luck and Vogel: brief displays of arrays of oriented coloured lines; suggests capacity limitation of ~4 objects

 

Verbal short-term memory

  • Rehearsal is necessary for long-term memory

    • Holding sequences of acoustic or speech-based items

  • The phonological loop was originally conceived as a phonological store + rehersal mechanism

  • Some say there is no phonological "store"

 

Articulatory suppression:

  • Silently mouthing words while performing a task

    • Impairs the phonological loop

 

Contemporary models of working memory

  • ^suggest there are no specialized short-term memory stores

  • Short term memory is the temp activation of long term memory

    • Frontoparietal attention and executive mechanisms access stored info

Background Visual Working Memory

  • What is it

    • Baddeley's model

 

  • Memory performance is worse with more items

    • Correct decreases as numbers of items increase

  • Continuous report delayed estimation

    • More errors are farther away from the color as you see more items

 

 

LO

  • What is novel about this research?

 

Delayed estimation of color (continuous report)

  • Color space is constructed to be perceptually uniform (color wheel is the same so they can measure exact error of color)

 

  • Bias: a systematic shift of responses away from the true value

 

Biases in delayed estimation of color

  • Stimulus-specific biases: there is a bias depending on the specific color

    • Memories for color are biased by prior knowledge/color categories

    • Small effect sizes make biases difficult to measure

 

The Current Work

  • Method: serial reproduction (iterated learning)

    • Long-term memory and linguistics to understand how prior knowledge influences memory

    • Used in visual working memory to study visuospatial priors

 

Why is this research novel?

  • Here we have each chain of responses measured within a single individual (rather than by different individuals) and have the starting value of each chain be the same across individuals

    • This lets us look at group level, item level, and individual level differences

 

Questions about this data

  • Group level: across participants and colors

  • Item-level: across participants, specific colors

  • Individual level: across colors, specific participants

 

Experiment 1: stimulus specific biases

  • Some colors are reported more frequently than other colors, even when the distribution of colors shown is uniform

  • There are biases that vary in magnitude and direction depending on the specific color

  • The iterated design amplifies this bias

 

What are group level properties ?

 

What are the item-level properties ?

 

Slope measures bias

 

Slope distribution show stimulus-specific variability

 

What are the individual level properties?

  • Different clusters

  • Most have at least 3 attractors

 

Experiment 1 conclusions

  • Inidivudal serial reproduction amplifies bias

    • Group level, item level, and individual level patterns

  • There are 3 group level "attractors" that responses drift toward across iteration

 

 

 

 

 

Psychology Long-Term memory

 

LO

  1. Memory is constructive, not a video camera

    1. Why might the brain prioritize encoding some things over others

 

 

Learning as "preparation for retrieval"

  • The best way to think about learning is as preparation for retrieval. Your mind is trying to predict when you'll need information again, and make it accessible then

  • This means memory is context based

 

Memory is context dependent (Baddeley)

  • Memorize something in the same location you learned it

  • Produces better recall

 

  • Mood also is important,

    • You perform better if test sad when you memorized something sad

 

But how does this help us?

 

What process best explains this?

  • Spreading activation: travels from one concept to another via associative links

  • Helps explain context effects, and also why partial information helps in retrieval

 

 

Spreading activation

  • This kind of activation spreading seems to happen automatically. Explains repetition priming

  • If words are connected, response time is faster

 

 

  • Items shown in similar conditions are recalled better than dissimilar conditions

LO

  1. What was the general experimental setup and conclusions made from Kandel's Aplysia experiments?

  2. What is a memory trace and how does it relate to the stages of LTM

  3. Be able to explain which brain regions are important for declarative memory and the different types of declarative memory

 

 

What is long-term memory (LTM)?

  • Stored information that does not need to be presently accessed or consciously accessible

    • Virtually unlimited capacity

  • All info from minutes, hours, days, and years ago is in LTM, unless its presently brought into working memory

 

Eric Kandel (Nobel Laureate)

  • Measured the gill-siphon-withdrawal reflex by recording single neurons in Aplysia California

  • Learning produces changes in behavior by modifying the strength of connections between nerve cells at the synapse

 

Aplysia as a model organism

Advantages

  • Relatively few neurons ~20,000

  • Large neurons seen with the naked eye

  • Easy to measure AP of single neurons and synapse structure

Implicit memory for reflexes

  • Habituation: decreased response to repeated exposure to stimulus

  • Sensitizing: enhanced response by giving an aversive stimulus

 

Stages of LTM

  • Memory trace: a bio change in the brain which reflects the storage of memory

    1. Encoding: creation of the memory trace

    2. Consolidation: stabilization of the memory trace

    3. Retrieval: accessing the LTM

 

2 categories of LTM

  • Declarative memory (explicit): memories that can be consciously accessed and, hence, can typically be declared

  • Non-declarative memory (implicit): memorie sthat cannot be consciously accessed

 

 

 

Medial Temporal Lobe (MTL)

  • Contains the primary structures which support encoding, consolidation, and retrieval

  1. Hippocampus

  2. Adjacent cortices

    1. Parahippocampal

    2. Entorhinal

    3. Perirhinal

 

Hippocampus

  • The hippocampus consists of subfields, some with more defined roles

    • Dentate gyrus

    • CA1

    • CA2

    • CA3

    • Subiculum

  • Damage to the hippocampus profoundly impairs declarative long-term memory

 

 The remembering brain: long-term memory

 

Diencephalon

  • Substructures of the diencephalon most implicated in memory

    1. Mammillary bodies of the hypothalamus

    2. Anterior and dorsal nuclei of the thalamus

  • Damage causes severe impairment in learning new info

    • Korsakoff syndrome (wet brain) (in severe alcoholism)

    • Stroke

 

Types of declarative LTM: episodic

  • Episodic memory: memory for an event that is tied to a specific time and place (what, when, where)

    • Autobiographical: an event in reference to yourself

    • Non-autobiographical: an event tied to a time and place but not in reference to yourself

    • Learning happens once at the time of the event (single trial learning)

  • Semantic memory: conceptually based knowledge about the world, including knowledge of people, places, and the meaning of objects and words

    • Facts

    • Schema/semantic memory structures

    • Personal semantics

    • Learning happens over repeated exposure across different contexts

 

Nondeclarative (implicit)

  • Skills are supported by basil ganglia

 

Types of longer term memory

  • Non-declarative memory: retrieval of info from long-term memory through performance rather than explicit/conscious remembering

  • Often learned gradually over time

    • Procedural- learning piano

      • Muscle memory for scales

      • Motor memory

    • Priming

    • Classical conditioning

    • Non-associative learning

 

Amnesia

  • Deficits in long term memory that cannot be explained by other cognitive deficits

    • Anterograde: impairment in memory for the 'future'

      • Cant retain new info

    • Retrograde: impairment in memory for the past

  • Observed in a variety of clinical disorders

    • Dementias

    • Brain injury

    • Medications

    • Drugs/alcohol

 

Case Study: patient H.M

  • After being knocked off a bike at age 9, he began to have seizures

  • Despite heavy and varied anti-epileptic meds, his seizures increased in frequency and severity

  • By age 29, he unable to work and had an extremely low quality of life due to seizures

 

  • Scoville offered to conduct an experimental surgery he carried out in psychotic patients

    • He would cut out parts of the temporal lobe, lobotomy

  • In 1953, and with his consent, HM underwent a temporal lobectomy to remove;

    • Both left and right hippocampi

    • Amygdala

    • Nearby medial temporal lobe cortex

HM Neuropsychological testing

  • Underwent neuropsychological testing in '55

  • On the day of the test, he said the date was march '53

  • Presented w/ intact IQ and cognition except for memory impairments

 

HM performance on memory tests

  • He could complete the digit span - not impaired on working memory

 

  • Star mirror trace test - procedural memory (learning skill or habit)

    • HM

      • First trial, many errors

      • w/ practice, his errors decreased

      • Motor skills intact!

      • Implicit memory spared

      • BUT, he had no memory of ever completing the tracing task before

        • His declarative memory was damaged

  • Medial temporal lobe and diencephalon damaged -> declarative memory damaged

  • Basal ganglia intact -> nondeclarative intact

 

  • Could complete semantic memory tests at a similar level of performance as controls

  • Temporally-graded retrograde amnesia

    • Could accurately remember remote episodic memories (age 16 and younger)

    • Could not remember episodic memories from 3 years before surgery to date of testing

 

 No Hippocampus

  • Impaired recent memory

  • Intact remote memory

  • What can we infer about the role of the hippocampus in declarative long-term memory retrieval?

    • Some responsibility for retaining long term memories

    • Consolidating memories: helps stabilize memories so you don't lose them

 

Temporally graded retrograde amnesia

  • Declarative memories acquired just before lesion (recent memory) are impaired

  • Declarative memories acquired in the distant past before the lesion (remote memory) are spared 4

  • Conclusion: the hippocampus is necessary for retrieving recent but not remote declarative memories -> evidence

  • The phenomenon of TGRA predicts theres a neural mechanism which stabilizes memories over time so that they are eventually remembered without the hippocampus

 

Long-term memory consolidation

  • Memory consolidation: the process by which a temporary, unstable memory trace is transformed into a more stable, long-lasting form

 

Systems Consolidation Theory

  • All declarative memory retrieval eventually becomes independent of the hippocampus

    • At encoding, the hippocampus and cortex are highly connected

    • As time pass after encoding, the memory reorganizes and becomes represented in the cortex

  • Systems consolidation theory suggests that the complexity distribution, and connections of the memory trace change with the age of the memory

 

Functional magnetic resonance imaging (fMRI)

  • MRI scanner

  • Processed fMRI data

    • Indicates which areas of the brain have more "functioning' (oxygenated blood) during an experiment

fMRI study of standard systems consolidation theory

 

Can the predictions of systems consolidation be observed for episodic memory in healthy adults?

 

Functional Connectivity

  • ^ decreased between the hippocampus and parahippocampal place area across 1 month

  •  functional increased between the vmPFC and the parietal lobe across 1 month

 

fmRI study of standard systems consolidation theory

  • They replicated their findings

    • Hippocampal brain activity decreased across 1 month

    • Frontal brain activity increased 1 month

 

Different Accounts of MTL and Memory

 

Is there a jennifer aniston neuron?

  • Researchers believ ethey discovered individual hippocampal neurons that code for a single, complex memory (grandmother cell)

 

Memory Coding Schemes

  • Local: a single hippocampal neuron codes for a single memory (grandmother cell)

  • Sparse distributed: each memory is coded by the activity of a small proportion of hippocampal neurons, and each neuron contributes to the representation of only a few memories

  • Fully distributed: each memory is coded by a pattern of activity across many neurons