Cognitive Neuroscience II
Chapter 8 Memory: Varieties and Mechanisms
Memory is the series of processes where the nervous system acquires
information, retains this information and uses it to guide behaviour and plan
future actions.
Memory consists of three phases: encoding, which consists of processes where
experiences can alter the nervous system, primarily through changes in strength
and/or number of connections between neurons. Storage, the retention of these
memory traces over time. Retrieval, the accessing of stored memory traces,
which may lead to a change in behaviour.
The term learning can refer to the encoding phase or to the combined efect of all
phases.
Working memory mediates the maintenance and manipulation of information for
a few seconds or minutes. Long-term memory is divided in declarative memory
(explicit memory), which refers to conscious memory for events and facts, and
nondeclarative memory, which refers to memories that are expressed through
task performance, independent of consciousness.
Amnesia due to brain damage most often arises from bilateral damage to the
medial temporal lobe. Unilateral damage produces relatively mild memory
defcits.
There is a double dissociation: damage to medial temporal lobe impairs
declarative but not working memory, while damage to the left temporoparietal
lobe does the opposite. The medial temporal lobe also impairs declarative but not
nondeclarative, what the occipital lobe does in reverse.
The major forms of nondeclarative memory are: Priming, the change in
processing of a stimulus due to a previous encounter of the same or related
stimulus. Skill learning, the gradual improvement in performance due to repeated
practice. Conditioning, drawing simple responses to associations between stimuli.
Priming
Priming can be classifed as direct priming, where prime and target stimuli are
the same, and indirect priming, where they are diferent (e.g. envelope -> letter,
semantic priming). Direct priming can further be classifed in perceptual priming
and conceptual priming.
In perceptual priming, the test cue and target are perceptually related (E_V_L_P_
-> envelope). It depends on brain systems diferent than those underlying
declarative memory. Patients with medial temporal lobe amnesia were impaired
in episodic memory, but not in perceptual priming. Conversely, patients with
occipital lobe damage were impaired in perceptual priming but not in episodic
memory. This suggests that perceptual priming depends on sensory regions of
the cortex.
Perceptual priming is not afected by levels-of-processing manipulation, but is by
study-test format shifts. For episodic memory, this is the opposite way around.
,Presenting diferent versions of the same object eliminates the repetition
suppression efect in the right fusiform gyrus but not in the left fusiform gyrus.
This result has been interpreted as evidence that the left fusiform gyrus stores
more abstract object representations.
The sharpening theory states that when a stimulus is repeated, neurons that
carry critical information about the stimulus continue fring, but neurons that
aren’t essential respond less and less. This causes reduced hemodynamic
responses.
Conceptual priming reflects prior processing of a stimulus’ conceptual aspects.
Conceptual priming is diferent from declarative memory, because it does not
depend on conscious awareness and is preserved in patients with medial
temporal lobe damage.
The fact that unlike amnesia patients, Alzheimer’s patients are impaired in
conceptual priming, suggests that conceptual priming is mediated by brain
regions afected by Alzheimer’s but not by amnesia (such as lateral temporal and
prefrontal cortices).
In semantic priming, the target word and prime are diferent but semantically
related. One theoretical idea of semantic priming is that the semantic memory is
organized in a network, and spreading activation causes semantic priming.
Skill Learning
Skill learning refers to gradual changes in behaviour due to extensive practice.
Skill learning is just like priming independent of the medial temporal lobes and
thus preserved in amnesia patients. It appears to rely heavily on the interaction
between the neocortex and subcortical structures.
The most important motor sequence learning task is the serial reaction time task
(SRT). This task involves four screen locations with four response keys. When a
stimulus appears at a location, subjects must press the corresponding key. The
order of stimulus location follows a repeated sequence. Results show that even
when a participant is unaware of such a sequence, sequence repetition leads to
faster reaction times.
Motor sequence learning tasks are impaired in patients with basal ganglia
disorders (such as Parkinson’s).
In the early learning phase (rapid improvement in performance), activation takes
place in the PMC, SMA, BS, cerebellum and parietal cortex. In the advanced
learning phase (slowly improving performance), activates specifc subregions of
the cerebellum and M1, reflecting the need for further error correction.
Perceptual skill learning refers to improvements in processing perceptual stimuli
that are identical or similar to stimuli that have been repeatedly encountered.
One way to examine perceptual skill learning is to train participants to recognize
Greebles. Researchers found that expertise in recognizing Greebles is associated
with increased activity in the FFA, which suggests that the FFA is related to visual
expertise rather than being a face-specifc module.
Cognitive skill learning refers to problem-solving tasks in which subjects are
required to use various cognitive skills to solve a task. One paradigm, the
weather prediction task, showed that patients with Parkinson’s disease never
, learned the task, where amnesia patients did. When the task was designed to be
an episodic memory task, results swapped.
Cognitive skill learning frst involves medial temporal regions, but these regions
are not necessary for subsequent learning of implicit rules.
Conditioning
Classical conditioning (Pavlov) involves an already present, innate reflex to be
modifed (dog sees food -> salivation to dog hears bell -> salivation). In operant
conditioning (instrumental learning, Thorndike, Skinner), the probability of a
behavioural response is altered by associating the response with a reward.
In delay conditioning, the CS is still present when the US starts, while in trace
conditioning, there is a brief time interval between the CS and US. In trace
conditioning, there must be a memory trace for a CS-US association to be
established, and this is not the case in delay conditioning.
Goal-directed actions are sensitive to action-outcome contingencies, whereas
stimulus-driven habits are controlled by stimulus-response associations. Goal-
directed A-O learning depends on the hippocampus and dorsomedial striatum
(caudate), whereas habitual S-R learning depends on the dorsolateral striatum
(putamen).
Donald Hebb frst hypothesized that memories are stored in the brain in the form
of networks of neurons that he called cell assemblies. Through experience,
assemblies then gradually become specifcally associated with objects, actions or
concepts.
Hebb proposed that when neurons are simultaneously active, the synaptic
connections between them are strengthened. This mechanism, called Hebbian
learning, operates via the ‘cells that fre together wire together’ adage. Hebb’s
proposals remain a plausible basis for understanding memory in cellular terms.
Habituation is a reduced response when the same stimulus is repeated,
sensitization is an increased response to the habituated stimulus when it is
paired with an aversive stimulus such as a shock.
Lomo and Bliss showed that when one pathway was stimulated with a high
frequency-stimulus, the postsynaptic neuron later showed stronger responses to
input from the stimulated pathway but not from others. This is called long-term
potentiation (LTP).
LTP has several properties that poses itself as an attractive candidate for one of
the major cellular mechanisms of memory:
- it can be induced by a single high-frequency stimulation (memory formed by
single experience)
- it can last for days or even weeks
- has a degree of specifcity (synapse-specifc)
- has a property of associativity
During the LTP, two parameters of the EPSP (excitatory post-synaptic potential)
change: the slope and the amplitude of the population spike (a measure of the
number of action potentials in the electrode area).
