Memory systems: Episodic

Goals

To complete our discussion of memory deficits following dlPFC lesions, and to describe the subcomponents of working memory.
To provide an overview of memory systems.
To present individual case studies of brain damage and memory deficits.
To discuss the medial temporal lobe and memory.
To discuss consolidation of memory

Topic slide

Henry Molaison (HM) (1926-2008) is a patient who had a bilateral resection (removal) of the hippocampus at age 27 after suffering from seizures due to a bicycle accident at age 5. He had a profound anterograde amnesia, a two-year retrograde amnesia, and a graded retrograde amnesia for the prior 11 years.

Brenda Milner (b. 1918) is an internationally acclaimed memory scientist in Montreal who studied HM’s memory deficits. Professor Milner still occasionally teaches at McGill University at age 101.

Reading

Reading: PN6 Chapter 30

dlPFC and memory

dlPFC lesions and memory

In my last lecture, we discussed two specific memory deficits associated with dlPFC lesions:

Temporal order and recency (did I see this item first?)
Source memory (where did I learn this?)

Can all of these deficits be related to a failure of executive control and/or allocation of attention?

Components of working memory

Dorsolateral prefrontal cortex plays an important role in working memory. Working memory is a short-term memory with a limited capacity in which information is maintained for short periods (~20 sec) – just long enough to recall for use (like a phone number).

Most researchers distinguish among phases of working memory – i.e., encoding, storage or maintenance, and retrieval.

I reviewed the intracranial ERP waveform from dlPFC during working memory task that I showed in last lecture – and pointed out phasic changes in the waveform that might be related to encoding, maintenance, and retrieval intervals

Encoding – the subsequent memory task

Another type of memory deficit is a failure to semantically organize when trying to recall a list of words. Most individuals will spontaneously report items organized by categories – such as reporting color words (red, green, orange). This suggests a deficit in encoding.

Encoding has been studied using the subsequent memory task using both fMRI and ERP studies of memory. Physiological data is acquired during encoding, encoding trials are sorted by later memory for items, average activity examined for ‘later remembered’ compared to ‘later forgotten’ trials.

This task was first used by Ken Paller (Northwestern University) with ERPs, where he described the 'difference due to memory', or Dm ERP phenomena. The ERPs showed a robust difference between later remembered and later forgotten words. However, the ERPs did not reveal where, in the brain, these ERPs were generated.

Anthony Wagner (Stanford) later used this same task with fMRI. He found that the dlPFC showed stronger activation for later remembered than later forgotten items. He also found content specificity by hemisphere:

Larger left dlPFC activation for remembered words
Larger right dlPFC activation for remembered faces.
Bilateral dlPFC activation for remembered objects.

Comparison of encoding and retrieval

I showed the meta-analysis by Cabeza and Nyberg that showed retrieval also activated PFC by hemisphere.

What does this mean? We have subsequent memory effects, retrieval effects, and working memory effects, all in the same dlPFC brain region. Returning to an earlier question: could this reflect a general effect of executive control and/or attention to internal stimuli ? Could dlPFC be communicating to sensory/perceptual regions to keep information 'on line' and, thus, playing an important, albeit indirect, role in memory?

Memory systems

Different aspects of memory.

I discussed the various meanings of memory – from episodes , to facts, to skills, to unconscious effects of priming:

Working memory: Holding a phone number in memory long enough to dial the number.
Episodic memory: Recalling an event from your past – complete with details about what, who, when, where.
Semantic memory: Your knowledge of who John F. Kennedy was, or what the word ‘cat’ means.
Procedural memory: Your knowledge of how to compute long division, play guitar, or throw a spiral.
Priming: Your speeded response when you see the word ‘cat’ presented a second time in a living/non-living discrimination test.
Conditioning: After pairing a tone with a puff of air to the air, you blink when you hear the tone.

Single dissociation of short-term (STM) and long-term memory (LTM)

Elizabeth Warrington and Tim Shallice studied patient K.F. in 1969 and demonstrated that K.F. had good LTM memory but abysmal STM (for auditory information, he had relatively preserved visual STM). He could not remember items when presented with short delays. However, he could learn paired-associates and could learn word lists, and recall them the next day as well as control subjects.

Are there dissociations between short and long-term memory? Compare Patients K.F. and H.M.

Another patient, E.E., studied by Hans Markowitz also shows STM and LTM dissociation – he can’t immediately recall a word list, but can do so after a long delay.

Different forms of long term memory

Long term memory is often grouped into two major categories:

Declarative memory or Explicit memory (those memories that can be stated or declared):
- Episodic (memory for events)
- Semantic (memory for knowledge)
Non-declarative or Implicit memory (those memories that cannot be stated or declared)
- Procedural (memory for skills)
- Priming
- Conditioning

Clinical case studies

Before beginning a detailed discussion of Declarative Memory, I discussed several clinical case studies.

Hyperthymesia

Subject ‘A.J’ could remember specific events from nearly every day of her life. Confirmed by diary entries. Found this memory to be burdensome in her life. Provided examples of her memory for each Easter over a decade. There are now 25 peer-reviewed case studies of individuals with hyperthymesia

Korsakoff syndrome

Korsakoff syndrome is due to thiamine deficiencies associated with chronic alcoholism. Petechial hemorrhages occur in the mammillary bodies, which are part of the main output of the hippocampus.

I discussed briefly the case of Jimmie G. – the ‘Lost Mariner’ in Oliver Sack’s book. Jimmie was a sailor in WWII who suffered from chronic alcoholism and was ‘stuck in time’ – thinking himself still a young man in the navy, although it was decades later. Chronic alcoholism irritates the gut, and reduces the absorption of thiamine. This causes brain damage, particularly in the mammillary bodies, which is a main output target of the hippocampus.

The video artist Tess Martin made an imaginative very short film based upon Jimmie G’s story which I have embedded below:

Herpes encephalitis

Clive Wearing (b. 1938) is an accomplished musician who developed a severe herpes simplex encephalitis that resulted in a severe anterograde memory (no ability to form new memories) and a substantial retrograde amnesia (loss of older episodic memories). I discussed several aspects of Wearing's memory loss, and illustrated the discussion with a compilation video made from several videos available on YouTube. If you find this case fascinating, try typing 'Clive Wearing' to the YouTube search bar.

His knowledge of his children was frozen in time (he remembered them as they were at the age of his illness) but he did know his wife. The video clips showed Wearing meeting his wife as though he had hadn’t seen her in years – twice within a short interval between. They also showed his diary entries – ‘awake for the first time in many, many weeks’. He lives on a tiny fragment of time with just void in the past.

I also showed a video clips of him playing music after he denied that he knew anything about music. Nevertheless, he could play piano and direct his choir when he was led to the piano. So his procedural or skill memory (piano playing) appeared intact.

One additional aspect of Wearing's deficit is semantic memory – no longer knew some things he used to know. He also learned almost nothing new since his illness. The psychologist identified about four items he appeared to have learned since his illness – e.g., reunification of Germany, and that Hong Kong was no longer a British protectorate.

I showed an autopsy brain (not Wearing’s, he alive at 81 y/o in 2019) illustrating the widespread damage caused by herpes encephalitis. It showed how the temporal lobes were particularly devastated, including the medial temporal lobe and, particularly, the hippocampus.

Oliver Sacks wrote an article for the New Yorker featuring Clive Wearing with the title "The Abyss". You can find it here

Declarative memory and the MTL

HM and RB

I described and discussed H.M.’s anterograde and retrograde amnesias, despite good short-term memory. The dissociation between long term and short term memory constitutes a single dissociation.

Patient R.B. was discussed as another example of severe anterograde and retrograde amnesia. R.B. had a loss of cells in one region of the hippocampus due to loss of oxygen during cardiac surgery.

The significance of R.B. is that his damage was far more restricted to the hippocampus than H.M. – who also had the most anterior parts of the temporal lobe removed so the surgeon could gain access to the hippocampus. Thus, it appears that the hippocampus is the critical structure.

Double dissociation of STM and LTM

Patient K.F. (with posterior brain damage – not hippocampus) shows very poor STM but good LTM.
Patient H.M. (with hippocampal resection) shows good STM but poor LTM.

Episodic memory in evolution

Up until this point, it has been easy to appreciate the adaptive advantage of sensory and motor systems, and the attentional systems that focus sensory resources and motor responses toward relevant stimuli in the environment. But what is the fitness advantage of having episodic memory? Are episodic memories unique to humans? While the conscious phenomenology of a warm remembrance may be difficult or impossible to study in non-humans. However, one might operationalize episodic memory, and then consider the possible consequences for fitness.

This issue was discussed in a 2013 PNAS paper by Allen and Fortin.

I reviewed and extended my prior discussion of episodic memory. In Lecture 1 we discussed birds who cache food have relatively larger hippocampi than birds that do not. Allen and Fortin operationalize episodic memory as having three components: what, where, and when. Data from scrub jays demonstrate that they know where food was cached, where it was cached, and when.

Allen and Fortin argue for a fitness advantage for episodic memory based upon an events-in-context framework. Quoting from their paper:

The events-in-context approach capitalizes on the fact that, in the episodic memory system, information about specific events is tied to the spatial, temporal, and other situational contexts in which they occurred … Based on this operational definition, demonstrations that animals can remember events in context … provided compelling evidence that core properties of episodic memory are present in nonhumans. This capacity is often termed episodic-like to emphasize that, whereas it does not address the phenomenological aspects associated with episodic memory in humans, it satisfies three key behavioral criteria … :

Content: The individual remembers information about the event (“what”) and its context of occurrence (e.g., “where” or “when” it happened).

Structure: The information about the event and its context is integrated in a single representation.

Flexibility: The memory can be expressed to support adaptive behavior in novel situations.

I then presented an overview of the hippocampal circuitry in humans, and showed that a similar circuitry exists in birds. I also showed a recent study of 54 bird species that reinforce an idea introduced in lecture 1 that the hippocampus of birds that cache food is larger than in birds that do not cache food.

In my review of the hippocampal circuitry, I emphasized the concept that what and where 'pathways' into the human hippocampus are segregated anatomically (where the 'when' comes from is not well known, although I alluded to the discovery of 'time' cells in the hippocampus).

I ended with a discussion of a clade plot created by Allen and Fortin that asked when episodic memory first evolved.

Spatial memory

The rat hippocampus has place cells – neurons that fire when the rat is in particular spatial locations. These cells will figure prominently in our discussion of Complementary Learning Systems in our next lecture.

H.M. can draw the spatial layout of the house to which he moved AFTER his surgery. How does he remember this? There is more posterior hippocampus remaining in H.M. than the surgeon’s notes indicated.

Taxi drivers in London

Studies by Eleanor Maguire and colleagues have used functional and structural imaging, and behavioral studies, to examine the effects of experience upon the structure and function of the hippocampus. Her work has take advantage of the vast spatial knowledge required to be a London taxi cab driver. Her work:

shows activity in posterior hippocampus when recalling routes in the PET scanner.
shows that taxi drivers have larger posterior hippocampi, and this correlates with number of years working as a taxi cab driver.
shows that taxi drivers have smaller anterior hippocampi, which also correlates with years working as a taxi cab driver.
shows that the smaller anterior hippocampus is related to deficits in verbal memory. This suggests that there is a cost to overdeveloping the posterior hippocampus.

Quoting from a study of Eleanor Maguire comparing taxi drivers to bus drivers:

In conclusion, licensed London taxi drivers show that humans have an amazing capacity to acquire and use for navigation a highly complex spatial representation of a large city. Our ﬁndings suggest that this is accompanied by greater gray matter volume in the posterior hippocampus. We speculate that there may be to be a price to pay for this expertise, however, as reﬂected in less gray matter volume elsewhere in the hippocampus, and a compromised ability to acquire or retrieve new visuo-spatial information. Longitudinal within-subjects studies are required to examine direct relationships between navigation expertise, changes in hippocampal volume, and their cognitive sequelae. In addition, future work using methods complementary to MRI and VBM may ascertain if gray matter volume differences are due to neuronal size, neurogenesis, or perhaps dendritic or axonal arborisation

Consolidation of memories

I discussed the role of the hippocampus in memory – why new memories unable to be formed, but old memories persist in many amnesiacs.

One theory is that the hippocampus is important to bind together information stored in other cortical regions that constitute the memory. However, over time (perhaps because the hippocampus ‘trains’ cortex), the hippocampal binding is no longer necessary, and the memory is maintained by cortical connections alone. Thus, the hippocampus is a ‘rapid learner’ while cortex is a ‘slow learner’. We will return to this idea in our discussion next time of Complementary Learning Systems.

Electroconvulsive shock therapy (ECT) produces a window of retrograde and anterograde amnesia that suggests the time course of this ‘binding’ phenomena. Memories 6 months prior to ECT were disturbed. Anterograde memories were disturbed for 2 months.

The Human Brain

Professor Gregory McCarthy

lecture2024-17