Home Research The Temporal Huddle
Activity in the medial temporal lobe predicts memory strength, whereas activity in the prefrontal cortex predicts recollection


(Mayu)Activity in right perirhinal cortex and in hippocampus bilaterally increased as a function of subsequent item memory strength. Could you speculate as to whether the perirhinal cortex and hippocampus are redundant or whether they are coding separate information? And if separate, what are the real-world implications?
(Larry Squire) We certainly suppose that the functions of the hippocampus and the perirhinal cortex are not redundant. There is a great deal of data supporting this claim. For example, in single-unit studies of recognition memory in monkeys and humans, the perirhinal cortex tends to respond in an item-specific way, while the hippocampus responds with a more abstract, match-mismatch signal. We do not find compelling, however, the idea that the perirhinal cortex and hippocampus respond to neat, dichotomous psychological constructs such as recollection and familiarity. Indeed, virtually all comparisons between these constructs have been confounded with memory strength (see our Nat. Rev. Neurosci 2007 review).

What are the temporal dynamics of recollection? Does a sense of familiarity precede source memory or are they simultaneous and parallel processes?
Familiarity is classically thought to be fast and effortless while recollection is slow and requires effort and reflection. Our own data don't say much on the subject due to the limited temporal resolution of the fMRI method.
When source memory strength was examined independently of item memory strength, only regions of prefrontal cortex and insula were identified. Should this activity be interpreted as having a specific role in memory per se, or is it task-related activity (i.e. greater activation in prefrontal areas when making an animacy or size judgment is correlated with "deeper" or "more effortful" processing, which in turn leads to subsequent better recall)? Is there any indication from reaction time data that longer RTs were associated with greater prefrontal (or hippocampal) activity during the encoding task?
We interpreted the frontal activity as reflecting enhanced processing of the stimuli during initial encoding that led to better, more confident source judgments at test. We didn't examine the relationship between RT and frontal activation. Interestingly, considerable neuropsychological data link frontal lobe damage to impaired source memory.
4) What is the current state of the debate about whether recollection and familiarity are independent processes that can be segregated functionally or whether the two effects in MTL regions are modulated by item memory strength? What would you predict would be the response of the Yonelinas' group to your findings?
The debate does not seem to be whether recollection and familiarity are distinct processes (they are), but rather whether that distinction can map onto anatomy (e.g., hippocampus vs. perirhinal cortex). Our conclusion in the paper was that much of the evidence taken to support this distinction between recollection and familiarity confounds memory strength with these processes. Our data indicate that when one controls for this confound, activity in both perirhinal cortex and the hippocampus responds to changes in item memory strength (familiarity). It is possible that the hippocampus and perirhinal cortex respond differently (nonlinearly) as memory strength increases (see our Nat. Rev. Neurosci., 2007, review Fig. 5). This is not to say that memory strength is the most important variable to think about in the case of medial temporal lobe structures (see point #1 above). And there may be other difference in their functionality as well. For example, anatomical considerations in non-human primates link perirhinal cortex to visual memory, and the hippocampus to a broader memory function. Skeptics might raise the possibility that our subjects did not trulyexhibit at-chance source memory in the condition when activity was calculated for trials when source memory performance was at chance. The argument is that subjects might have remembered some other details about the study episode, details that were not addressed by our source question. We consider this issue in the Discussion. There are also experimental approaches to this issue, which we are pursuing.