Temporal Memory And The Basal Ganglia

We followed up the memory failure hypothesis by investigating whether the deficits of PD patients result from dysfunction in either storing (writing to) or retrieving (reading from) temporal memory. PD patients were tested with a new design that withheld feedback on the second day of testing. Performance on the second day was therefore based solely on the retrieval of the information learned during the previous day. We used the standard PI timing procedure on the first day (the training session of the experiment) (Malapani et al., 1998a; 1998b; Rakitin et al., 1998). During the training session, patients learned and produced a short (6 sec) and a long (17 sec) interval, in separate blocks, with feedback provided in the same manner as in the previous study. On the following day (the testing session), subjects produced both intervals without further training and with no feedback.

We refer to this design as the encode-decode design (Malapani et al., 2002a; Rakitin et al., submitted). The key manipulation of the design in its initial form was the patients' drug state (ON or OFF medication) on the two successive days. PD patients were assigned to one of four experimental groups. The ON-ON group was provided with L-Dopa during both training and testing sessions. The OFF-OFF group was tested without L-Dopa for both sessions. The ON-OFF group was provided with L-Dopa during the training, but not the testing session. Finally, the OFF-ON group was tested without L-Dopa during the training, but with L-Dopa during the testing session. The idea was that by crossing PD patients' drug state with the availability of feedback, we could determine whether DA deficiency (associated with being in an OFF group) selectively affected memory storage (or encoding), retrieval (or decoding), or both, giving the design its name (Malapani et al., 2002a).

Two main effects are evident in these data. First, whenever subjects are OFF-drug, migration appears, replicating our earlier PD-related effect (Malapani et al., 1998b). That is, compared to the ON-drug state, PD patients tested OFF L-Dopa produce responses that are long compared to the 6-sec criterion and short compared to the 17-sec criterion. Migration was observed during both the training sessions, as seen in the OFF-OFF and OFF-ON groups' first-day performances, and the testing sessions, as seen in the OFF-OFF and ON-OFF groups' second-day performances. These results made it apparent that the migration effect was not dependent on the presence or absence of feedback, and hence was not likely attributable to a selective problem with memory encoding. This led us to attribute migration in PD patients to a DA-dependent dysfunction of retrieving temporal memories. A different type of timing deficit was observed in the OFF-ON groups' data from the testing session. In this case, patients substantially overproduced both target intervals. These data from the OFF-ON group contrast with the testing session data from the OFF-OFF group that clearly show strong migration. Apparently, restoring DA function during the testing session improves the retrieval deficit (manifested as migration) and allows a second, memory-encoding deficit associated with DA deficiency during the training session to be observed.

In addition to the direction of errors in accuracy, DA-dependent storage and retrieval distortions of temporal memories are also distinguishable by changes in timing variability. Psychophysical analysis of the data showed that the retrieval deficit, expressed with migration of productions, was associated with failure of the scalar property of timing variability. That is, the standard deviation of production response times — expressed as a proportion of the obtained, inaccurate means — is different for the two target intervals in all PD groups OFF L-Dopa, replicating our previous findings (Malapani et al., 1998b). The storage deficit, however, still reflects the scalar property of timing variability; that is, the standard deviation of response times rose in proportion to the magnitude of the timing errors. These results are illustrated in Figure 20.1 below.

To summarize, Figure 20.1 shows that DA deficiency leading to dysfunction in the basal ganglia results in two separable temporal memory deficits. Distortions that occur while storing multiple time intervals result in overestimation of all intervals during subsequent reproduction. In contrast, retrieving the trace of two or more different time intervals results in migration, a pattern of bidirectional errors such that reproductions of one of a pair of intervals are unusually close to the reproductions of the second. In addition to migration, DA-deficient retrieval is accompanied by a violation of the scalar property of timing variability, whereas with unidirectional shifts, the scalar property holds.

It is important to note here that estimation of continuous attributes other than time (e.g., line length) was spared in PD in an analogous encode-decode design (Malapani and Fairhurst, 2002; Malapani et al., 2002b). Moreover, patients with other kinds of brain damage, such as focal cerebellar lesions, did not show the temporal memory deficits seen in PD (Ivry, 1996; Malapani et al., 1998a, 2002a, b). Taken together, this evidence strongly suggests that the basal ganglia and their cortical targets are an important aspect of the neural basis of temporal memory in humans.

0 0

Post a comment