This is a fine article on the neurological syndrome of amnesia, with a small section on functional amnesia. I have only minor points and questions to bring up, and I’m happy to leave it up to the authors to decide how they’d like to respond.
It is clear that amnesia can arise from medial diencephalic damage, but it may be confusing that sometimes “medial” is left out (paragraphs 1 and 4).
The border between declarative memory and nondeclarative memory (paragraph 1) is perhaps presented as sharper than current findings allow, in that there are some memory expressions near the border that have not been easy to classify as fitting on one side or the other. This does come up in paragraph 22, but perhaps it would be helpful to mention at the outset that there are exceptions to the generalization (e.g., associative priming, trace conditioning). These phenomena may turn out to be theoretically important with regard to understanding the basis of the declarative/nondeclarative distinction, and leaving them out in the beginning sections may do more harm than good.
Material-specific deficits from unilateral brain damage are described as affecting either memory for verbal or nonverbal material (paragraph 2). Would it be more reasonable to describe these as relative deficits rather than absolute, given the strong evidence for bilateral cortical contributions to language and other functions?
In discussing the question of deficits in memory versus perception (paragraph 6), it should perhaps be added at the end of this paragraph that the quantitative measurements needed in neuropsychological analyses should not be limited to structural measures of brain integrity, because functional measures may show physiological disruption that is relevant to understanding the deficits and that is not apparent in structural measures. Also, Lee references are missing from reference list.
The term memory is used here to describe retrieved information, or the ability to make use of retrieved information, but one surprisingly different use is in the phrase “episodic memory stores spatial and temporal landmarks” which seems to shift to allowing “episodic memory” to be something that does the storage (paragraph 8). Another usage issue popped out later with amnesia (paragraph 15: “Amnesia impairs only long-term declarative memory”).
It might be helpful to clarify explicitly that the adjective “global” in “global amnesia” is meant to imply across all sensory modalities and domains, but not across all types of memory (paragraph 8: “global and encompasses” is unclear on this point, so this would be a good place for an explicit definition). The earlier use of “global” in paragraph 2 is unclear and probably unnecessary.
Saying that remote memory is “spared” (as phrased in paragraph 9) seems to imply that it is completely intact, whereas the general point is perhaps better to make here — that recent memory is more likely to be lost than remote memory, especially as all recent memories may not be lost.
I wonder how exactly the reader should interpret the concept of “the brain system that supports declarative memory” (paragraph 10), given that many regions of the cortex provide some support at various times for this ability, whereas this sentence could be (mis-)read as implying that the cerebral cortex does not support declarative memory. This point also pertains to a possible misunderstanding based on Figure 2, in that the brain regions listed do not comprise the sole or complete network required for the corresponding type of memory. The figure is problematic in that it appears to show one-to-one relations between each brain region and each type of memory, whereas these regions are generally only one part of the larger combination of networks involved. The problem may not entirely be cleared up by the later assertion (paragraph 12) that the neocortex is always important (and you could add there that the neocortex is important for memory retrieval early on).
The case for the idea that the hippocampal contribution ceases to be important is perhaps made too strongly with the statement in paragraph 12: “Memory is at that point supported by neocortex.” This is probably a little ambiguous and could be read as saying that it is always only the neocortex supporting retrieval of every remote memory, whereas perhaps the point to make is that such memories can be successfully retrieved if the neocortical substrates are intact even if the medial temporal substrates are not. I would not recommend that you state that the hippocampus is important for the whole lifetime of all episodic memories (as proposed by MTT), but my recommendation is perhaps slightly less controversial.
In discussing extensive retrograde amnesia (paragraph 10), it seems to me that it would be very useful to note that such deficits can arise from damage to cortical storage sites, from damage to cortical regions critical for retrieval processing, and/or from damage that disrupts consolidation.
The phrase “vulnerability of a short-term memory that has not yet been converted into a long-term memory” (paragraph 11) seems very unclear. The concept of short-term memory hasn’t been defined yet and contrasted with long-term memory, nor the idea of conversion from one to the other, which may be problematic anyway.
“Special” difficulty with tests of spatial memory (paragraph 13) is perhaps not as precise as saying that spatial memory was not found to be at the core of all memory deficits.
I find it interesting to speculate that normal conversations with amnesic patients depend not just on intact immediate memory (paragraph 16), but also on intact conceptual implicit memory for the topics under discussion. Later in this paragraph, it might be worth adding this one other reason for why the “difficulty for amnesic patients arises” as it is a key reason for explaining some otherwise anomalous findings — namely, difficulty for amnesic patients also arises when declarative retrieval of complex material is used to supplement the standard strategy of continuously rehearsing that information in working-memory tests.
How about replacing the phrase that nondeclarative memory “is non-conscious” (which is problematic in that the act itself, say, of reading repeated words, might be entirely conscious), with the idea that the awareness of retrieval need not (or does not) accompany performance (paragraph 17)? You then needn’t say that nondeclarative memory “operates outside of awareness” and you can move the “following examples” sentence to the end of the paragraph.
In the first skills paragraph, it seems too broad of a generalization to say that “amnesic patients can learn these skills at a normal rate.” I don’t know if any amnesics have been successfully taught to ride a bike, but I figure that most common motor skills (e.g., learning a backhand stroke in tennis, a basketball free-throw shot, how to play a new piece on the piano, a three-ball juggling pattern) are normally learned with a heavy declarative overlay of hints and strategies that are regularly recalled, and that such skills would not be learned at a normal rate by amnesics.
Priming — in addition to identifying or producing a stimulus, wouldn’t categorizing or evaluating a stimulus also be included in the definition? Whereas the first encounter may result in a representation of that item, perhaps sometimes it results in a change to an existing representation instead (though this distinction may or may not ultimately be useful). End of this paragraph: “poor memory of seeing the items earlier” could be changed to poor recall and recognition, given that implicit memory is not impaired.
Some terms are used as webpage links inappropriately: language in paragraph 5, neighborhood in the spatial memory section, attention in the immediate and working memory section, and Cs in paragraph 22, which of course shouldn’t link to the C programming language, and both letters should be capitalized (CS instead of Cs).