Motor Control: Memory and Motor Control in the Dorsal Striatum

Abstract

The dorsal striatum is important for motor control. Yet whether that control encompasses procedural memories, kinematic refinement, or both is still debated. A recent study has shed new light on the role of the dorsal striatum in learned movement sequences and the effort required to refine them.

To successfully navigate complex environments, animals rely on stereotyped sequences of movement that they have stored as procedural memories. For a person, motor sequences stored in this way may include those involved in riding a bike to work, ascending a staircase, and opening the door to an office. While these behaviors seem simple, they each involve a sequence of motor actions executed with precise timing and coordination. Procedural memories have been investigated for decades [1], but the location of these memories in the brain remains a point of debate [2,3]. One area that has been hypothesized to store procedural memories is the dorsal striatum: neurons in the dorsal striatum have firing patterns that may be sufficient for storing or recalling these memories, notably firing bursts that occur at different phases of motor sequences [4–6]. Manipulations of the dorsal striatum have been shown to impair the execution of procedural memories [7–9], suggesting that this structure is necessary for their storage or expression. The dorsal striatum has, however, also been linked to the control of movement kinematics: the coordination, timing, and magnitude of movement [10–12]. These two functions are difficult to tease apart in animal studies, as procedural memories are almost always expressed through movement. For example, if we lesion the striatum and find that an animal can no longer complete a sequence of lever presses, should we conclude that it forgot the procedure for completing the task? Or that it lacks the motor coordination to complete it? Behavioral assays are rarely designed to disambiguate these two possibilities. In this issue of Current Biology, Jurado-Parras et al. [13] report an elegant and creative approach that they developed to independently dissect the role of the dorsal striatum in procedural memory and motor control.

In the new study [13], rats were placed in a motorized treadmill where they could earn a drop of sucrose solution in a reward area at the front of the treadmill. The trick, however, was that they were required to wait 7 seconds after the start of each trial before entering the reward area or they would not receive the sucrose. In the first task, trials began with the treadmill belt moving slowly away from the reward area. To obtain the reward, most animals developed a stereotyped ‘wait-and-run’ strategy, where they would let the belt carry them to the rear wall of the treadmill, and would then run to the reward area to receive the sucrose (Figure 1A). Rats learned this strategy with close to optimal timing, taking slightly longer than 7 seconds to perform this sequence of actions. A different group of rats then learned a second task, which was identical to the first task except the belt now moved slowly towards the reward area. Most of these rats adopted a ‘run-and-wait’ strategy, wherein they ran to the back of the treadmill during the inter-trial interval before the trial started, and were then passively carried to the reward area by the treadmill belt (Figure 1B). The inter-trial interval was long enough that even mice with motor impairments could easily reach the back of the treadmill before the start of the trial. This ‘run-and-wait’ strategy also relied on animals not moving for most of the trial, thereby dissociating the procedural learning from motor kinematics. This was critical, as it meant that even animals with severe motor deficits could perform well on this task, so long as they could remember the procedure for completing it.

Figure 1. Schematic of the two treadmill tasks.

(A) Treadmill task wherein the belt is moving away from the reward and the animal was required to wait 7 seconds before accessing the reward. Most rats adopted a ‘wait-and-run’ strategy, where they were passively carried to the back of the treadmill by the belt, and then ran forward to get the reward. (B) Second variant of the treadmill task where the belt was moving towards the reward. Rats adopted a ‘run-and-wait’ strategy for success, where they first ran toward the back of the belt and were then passively transported forward to obtain the reward.

To test whether the dorsal striatum was necessary for either variant of the task, rats received fiber-sparing lesions to the dorsomedial striatum, dorsolateral striatum, or larger lesions that included both areas. In the first task, rats that received lesions to either the dorsomedial or dorsolateral striatum continued to use the ‘wait-and-run’ strategy, suggesting that neither area alone was necessary for this procedural memory. Rats that received large lesions that included both areas showed a deficit immediately after the lesion, seeming to forget the ‘wait and run’ procedure. But within a few sessions they recovered and began using this strategy again. These experiments suggest that the site of this ‘wait-and-run’ procedural memory is not in the dorsal striatum, or that it is only partially regulated by dorsal striatum.

Jurado-Parras et al. [13] acknowledge, however, that the immediate reduction in using the ‘wait-and-run’ strategy following the dorsal striatal lesion may reflect a deficit in procedural learning, and the animals may have re-learned this strategy de novo, possibly by new circuitry outside of the dorsal striatum. To address this, the authors employed the second variant of the task in which the direction of the treadmill belt was reversed. Here, they found that animals with lesions to the dorsal striatum kept using the ‘run-and-wait’ procedure for completing these trials, even immediately after the lesion. Together, these experiments argue against the dorsal striatum being a site of storage of procedural learning, at least for these two treadmill tasks. As animals completed similar numbers of trials after receiving the lesion, it further suggests that lesions to the dorsal striatum do not impair motivation to earn rewards.

In addition to providing insight into procedural memory, this elegant behavioral task also allowed for examination of the kinematics of movements on the treadmill. In contrast to the transient or non-observable deficits on procedural memory, running speed during the task immediately decreased and remained low following lesions of dorsal striatum. The magnitude of this decrease correlated with the size of the lesion, suggesting that the dorsal striatum is necessary for controlling movement speed or vigor. Critically, animals with lesions in the dorsal striatum were able to display a large range of running speeds when tested outside of the task on a treadmill set to various speeds. This confirms that their inability to move at high speeds in the rewarded tasks was not due to their inability to achieve those running speeds, but rather to a deficit in invigoration of the procedural memory during the task. The observation that dorsal striatal lesions reduced movement speed but had little influence on motivation or the strategy used to complete the task led Jurado-Parras et al. [13] to conclude that the dorsal striatum is critical for the tuning of motor kinematics but not the storage of procedural memory.

Finally, Jurado-Parras et al. [13] modelled these tasks by fitting optimal animal running behavior to a cost function that depended on the effort needed to maintain a given running speed. Predicted running trajectories were modulated by scaling a mathematical constant that weighted the contribution of running effort to the cost equation. Enhancing the weight of the ‘effort’ term in the cost function mimicked behavior of the animals after receiving lesions. This suggests that the behavioral strategies adopted after the lesions are less costly — running at lower speeds for longer periods of time requires less effort than a burst of vigorous speed during a shorter period of time. Cost modelling suggests that the role of the dorsal striatum in governing movement kinematics may lie in setting the sensitivity to effort, or in continuously influencing how animals execute vigorous actions [14]. Given the well-known role of dopamine transmission in striatal function [15], these results concur with the thrift hypothesis of dopamine [16], wherein striatal dopamine transmission regulates energy expenditure in response to environmental needs. In the first task, though it requires more energy to wait and then vigorously run toward the reward zone, this strategy is less likely to result in an error caused by prematurely entering the reward zone and forgoing the reward. Dorsal striatal lesions may distort how dopamine signaling regulates energy expenditure in pursuit of a goal, rendering animals more sensitive to effort. This interpretation is also supported by observations where Parkinson’s patients with bradykinesia displayed a shifted cost/benefit ratio for required movement speed, where they were less likely to move as quickly as controls, despite the fact that they were as accurate at every movement speed [17].

There are two limitations to the new work of Jurado-Parras et al. [13]. First, these studies left intact the ventral striatum, a brain structure which regulates goal-directed action, learning and effort [18,19]. It is possible that following dorsal striatal lesion, redundant circuits in the ventral striatum could have stored the procedural memories and enabled completion of these tasks. Second, while the authors demonstrated that procedural memory on these treadmill tasks does not require the dorsal striatum, it will take additional work to clarify the role of the dorsal striatum in other procedural memory tasks.

These results support a growing consensus that dorsal striatum is critical for modulating the vigor of movement [20]. They also argue against the dorsal striatum being necessary for storing procedural memories. This work raises important ongoing questions, including how to best think about the behavioral deficits in movement disorders such as Parkinson’s and Huntington’s, in which people experience deficits in both procedural memory and movement vigor. Are deficits in procedural memory in these patients due to deficits in motor kinematics? Or are the procedural memory deficits extra-striatal? Similar questions might be applied to compulsive disorders such as substance use disorder. These disorders are described more in terms of procedural learning and habit but are also linked to alterations in the dorsal striatum. Could changes in motor kinematics or effort underlie compulsive behavior? The Jurado-Parras et al. [13] paper is a testament to creative use of task design and novel behavioral analysis to address a long-standing unanswered question in the striatal literature. In sum, this work strongly supports that while dorsal striatal function does not underlie memory storage for the treadmill tasks presented here, it plays a significant role in the fine tuning of related movement vigor.