Adaptive control of thought-rational (ACT-R; J. R. Anderson & C. Lebiere, 1998) has evolved into a theory that consists of multiple modules but also explains how these modules are integrated to produce coherent cognition. The perceptual-motor modules, the goal module, and the declarative memory module are presented as examples of specialized systems in ACT-R. These modules are associated with distinct cortical regions. These modules place chunks in buffers where they can be detected by a production system that responds to patterns of information in the buffers. At any point in time, a single production rule is selected to respond to the current pattern. Subsymbolic processes serve to guide the selection of rules to fire as well as the internal operations of some modules. Much of learning involves tuning of these subsymbolic processes. A number of simple and complex empirical examples are described to illustrate how these modules function singly and in concert.

Single cells were recorded with 'tetrodes' in regions of the rat medial prefrontal cortex, including those which are targets of hippocampal afferents, while rats were performing three different behavioral tasks: (i) an eight-arm radial maze, spatial working memory task, (ii) a figure-eight track, delayed spatial alternation task, and (iii) a random food search task in a square chamber. Among 187 recorded units, very few exhibited any evidence of place-specific firing on any of the behavioral tasks, except to the extent that different spatial locations were related to distinct phases of the task. Furthermore, no prefrontal unit showed unambiguous spatially dependent delay activity that might mediate working memory for spatial locations. Rather, the cells exhibited diverse correlates that were generally associated with the behavioral requirements of performing the task. This included firing related to intertrial intervals, onset or end of trials, selection of specific arms on the eight-arm radial maze, delay periods, approach to or departure from goals, and selection of paths on the figure-eight track. Although a small number of cells showed similar behavioral correlates across tasks, the majority of cells showed no consistent correlate when recorded across multiple tasks. Furthermore, some units did not exhibit altered firing patterns in any of the three tasks, while others showed changes in firing that were not consistently related to specific behaviors or task components. These results are in agreement with previous lesion and behavioral studies in rats that suggest a prefrontal cortical role in encoding 'rules' (i.e. structural features) or behavioral sequences within a task but not in encoding allocentric spatial information. Given that the hippocampal projection to this cortical region is capable of undergoing LTP, our data lead to the hypothesis that the role of this projection is not to impose spatial representations upon prefrontal activity, but to provide a mechanism for learning the spatial context in which particular behaviors are appropriate.

Hippocampal cells that fire together during behaviour exhibit enhanced activity correlations during subsequent sleep, with some preservation of temporal order information. Thus, information reflecting experiences during behaviour is re-expressed in hippocampal circuits during subsequent 'offline' periods, as postulated by some theories of memory consolidation. If the hippocampus orchestrates the reinstatement of experience-specific activity patterns in the neocortex, as also postulated by such theories, then correlation patterns both within the neocortex and between hippocampus and neocortex should also re-emerge during sleep. Ensemble recordings were made in the posterior parietal neocortex, in CA1, and simultaneously in both areas, in seven rats. Each session involved an initial sleep episode (S1), behaviour on a simple maze (M), and subsequent sleep (S2). The ensemble activity-correlation structure within and between areas during S2 resembled that of M more closely than did the correlation pattern of S1. Temporal order (i.e. the asymmetry of the cross-correlogram) was also preserved within, but not between, structures. Thus, traces of recent experience are re-expressed in both hippocampal and neocortical circuits during sleep, and the representations in the two areas tend to correspond to the same experience. The poorer preservation of temporal firing biases between neurons in the different regions may reflect the less direct synaptic coupling between regions than within them. Alternatively, it could result from a shift, between behavioural states, in the relative dominance relations in the corticohippocampal dialogue. Between-structure order will be disrupted, for example, if, during behaviour, neocortical patterns tend to drive corresponding hippocampal patterns, whereas during sleep the reverse occurs. This possibility remains to be investigated.

The methodologies of cognitive architectures and functional magnetic resonance imaging can mutually inform each other. For example, four modules of the ACT-R (adaptive control of thought - rational) cognitive architecture have been associated with four brain regions that are active in complex tasks. Activity in a lateral inferior prefrontal region reflects retrieval of information in a declarative module; activity in a posterior parietal region reflects changes to problem representations in an imaginal module; activity in the anterior cingulate cortex reflects the updates of control information in a goal module; and activity in the caudate nucleus reflects execution of productions in a procedural module. Differential patterns of activation in such central regions can reveal the time course of different components of complex cognition.