Temporal And Sequential Control Interactions

The proposition that interval timing is an important function of the striatum is compatible with its role in other behavioral structuring functions. Striatal involvement in sequence processing has been demonstrated in studies with rats (Aldridge and Berridge, 1998; Berridge and Fentress, 1987), monkeys (Kermadi and Joseph, 1995; Miyachi et al., 1997; Shidara et al., 1998), and humans (Rauch et al., 1997; Toni et al., 1998). Like the control of behavior proposed for interval timing, efficient sequence processing involves directing behavior and attention to sequential elements of the target. Similar to our proposal for distributed striatal processing of time, sequence-related processing has been found across multiple striatal locations and appears to depend on the type of sequence processed and the degree to which the sequence execution has become automatic (Miyachi et al., 1997).

Given the role of the striatum in both sequencing and interval timing, it is not surprising that these two processes are closely related. This phenomenon can be seen anecdotally when humans are asked to time an interval, as they immediately begin counting (i.e., they utilize a sequential behavior strategy). The interrelation of these two processes can be further appreciated given the formulation of models such as BET (Killeen et al., 1988), where the temporal control of behavior does not result from a temporal percept, but instead from the sequential behavior produced by the organism. In contrast to relying on such sequential behavior for timing, we have proposed that sequential-temporal behavioral interactions may develop naturally due to the interacting processes of timing. The converse sequential-temporal interactions are also seen in models of sequence processing that rely on temporal information (Dominey and Boussaoud, 1997), and such an interaction is supported by evidence showing that disrupting the temporal relationship of sequential elements impairs sequence recognition (Dominey, 1998). Extrapolating from these interacting timing and sequencing examples, we conclude that the striatum may be involved in composing or directing behavior by utilizing numerous sources of information so that behavior is executed in an efficient manner. In other words, the striatum may act to guide behavior by controlling the when, which, and how much of motor and cognitive activity.


Although the SBF model successfully predicted the peak-type striatal data described above, we found no evidence of oscillating cortical cells, which were proposed to serve as the clock signal in the model (Matell and Meck, 2000, in press). Furthermore, although there is frequent mention of oscillatory processes in the recording literature (Ahissar and Vaadia, 1990; Gray et al., 1989; Steriade, 1997), cortical oscillations in an appropriate range are most frequently found during inactivity or rest, rather than when the animal is actively behaving — which would be necessary given the existence of the behavioral chains described above. As such, we feel that although the global framework of the SBF model quite likely describes timing in the brain, the reliance on cortical oscillations is a potential weakness of the model. We have not yet fully explored the effects of feedback loops on the generation of a peak output in our simulations. However, given that our model produced peak output without feedback, it is likely that adding feedback would disrupt its output. Similar disruptive effects may also be seen in other timing models following addition of feedback (although in some respects, BET is built on feedback). As such, we have begun to explore whether a cortico-striatal coincidence detection circuit coupled with a feedback loop can produce peak output without the reliance on a variety of oscillatory clock signals. Although we have not yet achieved an acceptable degree of similarity to the decision stage peak output, our results have been encouraging, and we intend to pursue these mechanisms further. Given the known circuitry of the cortico-striatal system, we believe that the investigation of feedback loops in other timing models will be a fruitful source of exploration.


Like our main thesis, we return to our initial statements in which we have outlined the logical possibility that interval timing in the brain might not be composed of independent, serial mechanisms, but rather operate via feedback loops in which the output of an internal clock serves to dynamically modify its own input. We argued that reference to a generalized timing model incorporating both input and output processes could provide guidance in the search for the neural mechanisms of timing because such processes likely play an important part in the production of a temporal percept. We have shown that striatal neurons can indicate times of expected reward through fluctuations in their firing rates, thereby suggesting a prominent role of the striatum in timing and time perception. Thus, it may be possible that the output of individual striatal neurons serves as a decision stage function of an interval timer. Furthermore, based on both our own recording data and anatomical evidence, it appears that the striatal timers are distributed throughout all functional areas of the striatum and are thereby embedded within regions organized for other behavioral and cognitive functions. Such an embedded interval timing system would be an efficient mechanism of organization for a process that is so heavily involved in every facet of behavior. Further, the distributed nature of the timing system may be an important contributor to the behavioral sequences that are engrained in the behavioral expression of temporal perception. Finally, we believe that a thorough investigation of the effects of feedback processes within interval timing models may steer the field of timing and time perception in promising new directions.

0 0

Post a comment