Review article
Temporal lobe epilepsy as a model to understand human memory: The distinction between explicit and implicit memory
The historical process of classifying memory systems such as explicit and implicit memory has been challenging, as it becomes increasingly difficult to place new data at either end of the dichotomy. In a recent review, Squire proposes that the debates over characterizing memory systems, such as the explicit/implicit distinction, are rooted in a more universal scientific mission to understand more clearly how exactly the brain acquires, stores, and retrieves new information [52]. Work conducted in patient populations has supported the general idea that there are different forms of consciousness that are differentially affected by disease or damage. What predicts specific impairments, however, has been perhaps the most puzzling, as one group of patients with MTL damage can often perform quite differently than another group on virtually identical experimental tasks [3,53,54]. One hypothesis to explain such differences suggests that performance is based, at least in part, on the severity and extent of damage of MTL structures [22]. This idea was made especially plausible in a review conducted by Gooding et al. in 2000, which found that in amnesic patients, severity of impairment on explicit and novel implicit memory measures was directly related to the extent of lesion [22]. An alternative, or perhaps additional, explanation relates to the type of material involved when assessing explicit and implicit memory for novel versus preexisting information. Yang et al. reported deficits only on novel implicit tasks that required forming new associations in memory [26]. In their study, patients who sustained damage to MTL regions actually performed normally on tasks that indirectly assessed learning and memory for single-item information (such as a single word), but demonstrated impaired performance on measures of both explicit and implicit memory for new associations between words. A more recent study reported impaired explicit and implicit memory for novel between-word associations in patients with MTL amnesia and, importantly, found that degree of priming correlated directly with the degree of episodic memory impairment (established through performance on standardized neuropsychological memory tests) [55]. This general finding has been documented elsewhere [48], leading to the conclusion that to-be-remembered material requiring an associative or relational component is more vulnerable to the effects of MTL damage, regardless of the direct (explicit) or indirect (implicit) nature of the task. In fact, the role of the hippocampus in relational aspects of memory is a dominant theme in contemporary memory research [56–58], and recent cognitive–neuroanatomic studies have focused on the hippocampus as critical for encoding relational aspects of incoming information, regardless of the explicit or implicit nature of the task. However, this idea has been refuted in the literature. Rosenbaum et al. reported that the amnesic patient K.C., who has extensive bilateral hippocampal loss, was able to demonstrate implicit associative priming for new associations [59]. The authors argued that MTL structures external to the hippocampus can be involved in the binding together of incoming, unrelated information. Despite this, it is also important to note that some amnesic subjects, such as K.C., may have adapted to their deficits over time and, thus, have essentially “learned” to use extrahippocampal brain structures when engaging in experimental memory tasks. Although it is certainly not “normal,” it nonetheless is enough to support certain types of memory encoding and retrieval. Thus, the degree to which hippocampal function contributes to performance on a memory task depends on a multitude of factors, including type of material to be learned or remembered [22], severity of damage [60,61], duration of disease or impairment [62,63], and which remaining brain structures are available [55,59].