Saturday, March 05, 2005

Au, A., Chan, A., & Chiu, H. Conceptual organization in Alzheimer's dementia Journal of clinical and experimental neuropsychology 25 (6): 737-750.

In this study of Chinese AD patients, the subjects were given two tasks. The first involved a triadic comparison: the patients were given a series of target drawings, each of which had two drawings accompanying it. One of the two companion drawings would be related in one of the following ways:
  • Taxonomic (i.e. the target is a picture of a carrot, and the associated picture is of a tomato)
  • Thematic (i.e. the target is a carrot, and the associated picture is a rabbit)
  • Shape (i.e. the target is a carrot, and the associated picture is a rocket)
The second task involved a more guided version, assessing their ability to make thematic, taxonomic and shape associations.

The findings:
AD patients tended to used perceptual information to categorize objects, and this tendency increases with dementia. "These findings were consistent with the results of the multidimensional scaling studies suggesting that AD patients in the West tended to rely heavily on the concrete perceptual dimension in categorization rather than the more semantically-demanding abstract conceptual dimension used by normal controls." (748)

The AD patients fared poorly on the thematic relationships: "By definition, thematic organization is based on spatio-temporal schemas and cause-effect relationships, thus relying heavily on the knowledge of specific attributes of objects. Therefore, the poor performance on the thematic task of AD patients in the present study may be interpreted as a loss of subordinate knowledge about the specific attributes of a semantic category." (748)

"The AD paitent group as compared to the control group chose significantly more shape objects in the Unguided condition though they made fewer correct choices than the controls in the Guided shape task." (748)
Ashby, F.G. & O'Brien, J.B. 2005. Category learning and multiple memory systems Trends in cognitive sciences 9(2): 83-89.

Ashby and O'Brien provide another helpful taxonomy, here:

Declarative memories: "those accessible to conscious awareness." These include:

  • Working memory: the ability to maintain and manipulate limited amounts of information during brief periods of cognitive activity.
  • Episodic memory: memories of specific past events or episodes in our personal history. This memory tends to be "context-rich," in that it includes information from all sensory modalities.
  • Semantic memory: memory for facts. This memory tends to be "context-poor," relying on only one modality. Presumably, both episodic and semantic memory are governed by the medial temporal lobes, particularly hippocampal and parahippocampal structures.
Non-declarative memory systems:

  • Procedural memory: memories of skills that are learned through practice, such as motor skills in sports. These memories involve little conscious recollection or even awareness; they are acquired slowly and incrementally; suspected to be mediated through the basal ganglia.
  • Perceptual representation memory system: in which repeated presentations of the same stimulus in a brief interval stiumulates that part of the visual cortical unit that is sensitive to that stimulus.
Keri, S., Kalman, J., Kelemen, O., Benedek, G., & Janka, Z. 2001. Are Alzheimer's disease patients able to learn visual prototypes? Neuropsychologia 39: 1218-1223.

In a follow-up to their earlier work, the authors describe a study of AD patients and their ability to learn visual prototypes. AD patients, when given the tasks of recognizing and categorizing visual dot patterns scored significantly lower than the control group in recognizing visual dot patterns. However, their ability to categorize was relatively spared.

The conclusions: explicit memory is markedly affected even in early AD (they call it the medio-temporal/diencephalic explicit memory system).

The sensory neocortical areas mediating Implicit category learning, on the other hand, appear to be spared in the early stages of AD.

Keri, S., Kalman, J., Rapcsak, S., Antal, A., Benedek, G. & Janka, Z. (1999). Classification learning in Alzheimer’s disease. Brain 1999; 122.6: 1063-1068.

There are a number of useful things here:

1. The distinction between explicit and implicit memory. Explicit memory involves a direct querying of the brain to receive an answer: e.g. “What is the capital of Canada? Ottawa.” Implicit memory is more closely related to procedural memory. The research of Squire and Knowlton with amnesic patients has suggested that prototype-based category learning is linked to implicit memory, rather than to explicit memory.

Alzheimer’s patients in this study, predictably, fared considerably worse than the control group in explicit memory tasks. But what about the category learning?

2. The distinction between prototype and exemplar-based models of categorization. In the prototype method, the subject forms a prototype from different exemplars, in a process similar to the formation of Frye’s archetype or Plato’s forms. Categorization consists of grouping exemplars with the prototype which they most closely resemble.

Exemplar-based learning, on the other hand, suggests a more emergent process, in which categories consist of groups of distinct exemplars, but without the representation of a single prototype.

In the study, Alzheimer’s patients fared considerably worse than the control group in prototype methods of categorization, but actually fared as well, if not better than the control group in exemplar-based methods of categorization. (They were less fazed by high distortions, for example.)

The findings suggest that the controls used prototyping and that the AD participants used exemplar-based methods.

3. The authors propose a 3-level model of category induction:

a. Neuronal representations of individual patterns are formed;

b. Patterns are grouped according to exemplar-based categorization;

c. In a subsequent phase of integration, a prototype is extracted from those categories.

If so, the findings of the study suggest:

a. That declarative memory suffers a robust impairment in AD patients (no kidding);

b. That categorization can carry on in the absence of a strong explicit, declarative memory;

c. That AD patients suffer significant impairment of prototyping categorization, but not exemplar-based categorization.

Weingartner, Herbert J., Dawas, Claudia, Rawlings, Robert, & Shapiro, Martha. 1993. Changes in semantic memory in early stage Alzheimer's disease patients. The gerontologist. 33(5): 637-642.

According to this article, changes in the semantic memory is not the most obvious symptom of AD: usually, we notice AD through changes in recent, or episodic memory. However, changes in semantic memory, while more subtle, are a more reliable indicator of Alzheimer's Diseases. This study examined the ability of AD patients to generate words that are exemplars of closed, or limited categories (such as fruits and vegetables) and their ability to generate words for open categories (think of a word that begins with "F," for instance).

They found that changes in semantic memory appear very early in the disease.

The progression:

a. Unable to generate infrequent, low-probability items that belong to categories of knowledge.
b. Later, unable to generate the more common and obvious elements that make up networds of knowledge.

These changes appear to be caused, not by a failure to retrieve available information, but by a loss of stored information.
Done, D. John & Gale, Tim M. 1997. Attribute verification in dementia of Alzheimer type: evidence for the preservation of distributed concept knowledge. Cognitive neuropsychology 14(4): 547-571.

Arguing for a neural network model of semantic memory, the authors present a study in which Alzheimer's patients were subjected to knowledge probe questions. The subjects revealed greater success in categorizing at the superordinate level than at the item-specific level.

"Despite a wealth of evidence suggesting that superordinate knowledge is more robust to cortical damage in dAT, this finding is in some ways counterintuitive: Evidence from normal subjects suggests that the basic level is the preferred level in spontaneous naming tasks, and it is also thought to be the level of categorisation that is first acquired by young childre, followed by the superordinate and then the subordinate level." (549)

Tuesday, January 18, 2005

Manulife Walk for Memories!

In honour of all who are struggling with this disease, I will be walking in the Toronto Manulife Walk for Memories.

May the future hold the answer to our dreams and prayers.

Saturday, March 13, 2004

I had no idea that there were so many different kinds of memory! I’ve been reading The Psychology of Dementia by Edgar Miller and Robin Morris, and they provide some very useful models from cognitive psychology that help to classify and category the various kinds of memory and memory functions.

Episodic memory: memory for personally-experienced events or material

Semantic memory: memory for information about the world, including rules, concepts and the meaning of words

Implicit memory: cognitive processing based on experience but without conscious retrieval of information

Skill learning: the development of perceptual motor skills

Episodic Memory

Short-term memory: recall of material or events after a period of up to 30 seconds

This kind of memory has a much smaller capacity, and often involves active rehearsal mechanisms (such as articulatory rehearsal). It is often tested by asking the subject to repeat a sequence of items, such as words or letters. AD patients frequently show substantial deficit in their ability to remember 3 words, especially if a delay has been introduced.

What exactly is happening?

Martin et al (1985) suggest that the verbal material is being encoded incorrectly in the patient’s brain, along with a language impairment: the stimulus attributes of the material are not sufficiently processed to produce a strong memory trace.

Morris & Baddeley (1988) suggest that short-term memory is divided into subsystems that are distributed across different domains of information processing, all controlled by a central controller that ensures that the systems operate in sequence and in a coordinated fashion.

Long-term Memory: memory lasting for longer periods, minutes, days or weeks. This stores a vast amount of information, but relies on more complex retrieval mechanisms.

Martin, A., Browuwers, P., Cox, C. & Fedio, P. (1985). On the nature of the verbal memory deficits in Alzheimer's disease. Brain and Language, 25, 323-341.

Miller, E. & Morris, R. (1993) The psychology of dementia. Chichester: Wiley.

Morris, R.G. & Baddeley, A.D. (1988). Primary and working memory in Alzheimer-type dementia. Journal of clinical and experimental neuropsychology, 10, 279-296.

Wednesday, March 10, 2004

Further thoughts, inspired by a wonderful conversation I've just had with a dear friend.

I've been reading further in books: combinations of neuropsychological literature about the cognitive dimensions of dementia, information about the history of Azheimer's research, and accounts of caregiving experiences, together with the blogs of people with Alzheimer's. Some things are emerging:

1. That there's a quality of engagement with life which seems relatively resistant to dementia: a reaction, for instance, to a beautiful landscape, the colours of trees, the sunlight on water, the warmth of summer. The sensations of comfort. These are feelings which are very much "in the moment": you don't need a lot of complex mental modelling to be able to see, feel, hear and appreciate these things.

2. That communication between a patient and caregiver is a matter of working hard to find a mental model that fits with the patient. Understanding how cognition changes under the effects of AD, how it alters peoples' methods of labeling, articulating, categorizing phenomena that they encounter, can only help the caregiver understand where the patient is in these circumstances.

3. People with Alzheimer's have much to teach us. We are obsessed, as a culture, with storing, organizing and accessing experience for the future. My friend commented on seeing someone in a beatiful French cathedral madly recording everything on the videocam, rather than simply standing there and taking it in. When Alzheimer's occurs, your ability to store things up to use at your convenience is greatly reduced. You can't remember things the way you used to; you can only experience them as they're happening.

A lot of this may be simple and naive. But it helps, nonetheless.

Monday, March 08, 2004

The Alzheimer Society has an interesting web page:

This provides links to weblogs of people with Alzheimer's. It sounds like the world's most depressing site, but actually, it isn't. If anything, it's a tribute to the monumental courage and resourcefulness of these people, who not only have Alzheimer's but are willing to share their experiences.

It also convinces me of an idea that's been growing in me for some time: that if we can maximize the time of meaningful communication between people with Alzheimer's and their caregivers, we will be doing a huge service, not just to patients, but to the caregivers. People with Alzheimer's are uniquely placed, at least in the early stages, to provide insights that are badly needed for the rest of us. The courage and humour and faith that shines through those weblogs is something that does me good, and would do many good in this frantic world.

Saturday, March 06, 2004

Canadian Memantine Trials

According to the Alzheimer Society of Canada, participants are now being recruited for clinical trials of memantine in Canada.

The ad is at:

They don't date the ad, which is kind of annoying, but the menu list indicates that the ad is new, February 2004.
Follow-up on Memantine

The January 2004 issue of JAMA has the following:

JAMA. 2004 Jan 21;291(3):317-24. Related Articles, Links

Memantine treatment in patients with moderate to severe Alzheimer disease already receiving donepezil: a randomized controlled trial.

Tariot PN, Farlow MR, Grossberg GT, Graham SM, McDonald S, Gergel I; Memantine Study Group.

According to this study, moderate to severe AD patients taking both Memantine AND Aricept had significantly better cognition, daily activities and behaviour. This is promising.

In the meantime, finding out about whether Memantine will ever be an option in Canada is proving to be a lot more difficult!

Done, D. John & Gale, Tim M. (1997). Attribute verification in dementia of Alzheimer type: evidence for the preservation of distributed concept knowledge. Cognitive neuropsychology 14(4): 547-571.

This is a provocative article, and one that has been challenged since its publication. It deals with categorization abilities in people with dementia of Alzheimer type (DAT), and suggests something that is counter-intuitive.

Since the 1970s, we have been profoundly influenced by Eleanor Rosch and others, who have argued for the existence of a "base level of categorization": a level which is linguistically simple, learned early, and, by implication, the last level to leave us. A child, for instance, will learn the word "dog," and then, later, will learn that dogs and cats are both mammals (a superordinate level), and that dogs can subdivide into beagles and terriers (subordinate level).

The evidence of the study described by Done and Gail suggests that the patients had better superordinate knowledge; it therefore supports "a neural network model of semantic memory, in which a concept is represented by a distributed pattern of activity across a set of features. In such models, there is no explicit distinction between basic level, subordinate level, and superordinate level features...." (from the Abstract).

I find this interesting, even though subsequent researchers have suggested that the variations in the findings were more due to the nature of the testing than intrinsic results. It suggests that there are practical limits to the notion of base levels of categorization, and that those levels of categorization are very important for our typical understanding. This is borne out by accounts from the caregiving literature, in which a patient will ask for a peanut for breakfast, when he means grapefruit.

If we could spot some patterns in the breakdown of categorization skills, we might be able to find ways for caregivers to communicate with Alzheimer's patients more effectively, and for longer periods.

Thursday, April 03, 2003

According to a recent article in the Washington Post by Shankar Vedantam (April 3, 2003), a drug called Memantine may be approved by the FDA in the US, and may have a significant effect in reducing the effects of moderate to severe dementia. The report of a controlled clinical study has recently been published in the New England Journal of Medicine. In this study, 252 patients with advanced Alzheimer's were treated over a 28-week period: those receiving Memantine deteriorated half as much as those treated with the placebo.

A separate study by Martin Farlow, a professor of neurology at Indiana University, suggested that Memantine, when combined with Aricept, actually caused some improvement. This report has been neither peer reviewed or published at this point.

We live in hope.