Motor and non-motor sequence learning in patients with basal ganglia lesions: the case of serial reaction time (SRT)

Abstract

In order to address the question of whether the basal ganglia are involved exclusively in regulation of motor sequence learning, or if they are involved in non-motor sequence learning as well, two versions of the serial reaction time (SRT) task were administered: First is the standard version of the SRT task in which the sequence is executed motorically, and the second is a non-motor version of the task which requires response only to a particular position of the sequence. Sixteen patients with damage restricted to the region of the basal ganglia and 16 matched control subjects participated in this study. In addition to the motor and non-motor SRT tasks, two declarative memory tests (Visual Paired Associates and Rey Auditory-Verbal Learning Test) were administered to the participants. Results indicate that the two groups did not differ either on learning rate of the two declarative tasks, or on the declarative component of the SRT tasks (i.e., ‘generate’). However, the control group was significantly superior to the basal ganglia (BG) group in learning a specific sequence in the motor and non-motor SRT tasks. Results suggest that the basal ganglia are involved in the regulation of non- motor as well as motor sequence learning.

Introduction

It is well established that the learning and memory of some forms of knowledge are preserved in amnesic patients [19]. Based on this notion, Cohen and Squire [2] have introduced the distinction between two forms of memory—‘declarative’ and ‘procedural’ memory. The former is memory for facts and events, while the latter is the ability to acquire and retain new skills. Declarative memory, which is impaired in amnesics, is typically tested by methods of recall and recognition. Procedural memory, which is preserved in amnesics, is tested by using a wide variety of tasks such as the Tower of Hanoi puzzle [3], mirror reading [2], and serial reaction time (SRT) [14].

While the role of the medial temporal and diencephalic structures in declarative memory is well documented, the brain structures subserving procedural memory are not yet as clear [20]. Some researchers have pointed to the basal ganglia as the crucial area for processing of procedural information [18]. In animal studies, damage to the basal ganglia has been found to affect procedural memory [12], but studies of patients suffering from degenerative diseases of the basal ganglia such as Parkinson’s disease (PD) and Huntington’s disease (HD), are less conclusive. PD patients were found to be impaired in a variety of skill learning tasks, such as complex tracking [5], SRT [4], [9] and the Tower of Toronto [17]. However, other studies do not support the basal ganglia hypothesis of procedural memory. Heindel, et al. [7] tested two groups of PD patients, one demented and the other not demented, with no difference between groups in terms of motor symptoms; they found that the patients’ impairment on learning the pursuit-rotor task was correlated with the degree of dementia but not with the severity of motor symptoms. Contrary to findings by Saint-Cyr et al. [17], in two other studies PD patients’ performance on a Tower puzzle did not differ from normal controls [1], [13].

Several attempts have been made in the literature to resolve these conflicting findings. Owen et al. [15] showed that PD patients’ performance is a function of clinical disability and precise index of performance (i.e., accuracy vs latency). Some researchers raised the possibility that the heterogeneity of PD patients has further contributed to inconsistent reports in the literature. Vakil and Herishanu-Naaman [23] found that the PD patients with bradykinesia, but not those patients with tremor as the predominant symptom, demonstrated impaired procedural learning. Other researchers have emphasized the heterogeneity of the procedural tasks. Harrington et al. [6] found that PD patients were impaired in the acquisition of a motor (i.e., rotary pursuit) task but not in the visual-perceptual (i.e., mirror reading) task.

Sequence learning as measured by the SRT task has been consistently shown to be impaired in PD patients [4], [9], [16] and in HD patients [10], [26]. These results were interpreted as support for the hypothesis that the basal ganglia are involved in the regulation of procedural memory or at least in sequence learning. In light of the findings by Harrington et al. [6], it could be argued that the impaired performance on the SRT task in PD and HD patients is due primarily to the fact that this task requires motor sequence learning. Thus, impaired performance on the SRT task could not be interpreted as reflecting impaired procedural or even sequence learning in general, but only motor sequence learning.

In order to address the question whether the basal ganglia are involved exclusively in the regulation of motor sequence learning, or in non-motor sequence learning as well, two versions of the SRT task were administered. The first is the standard version of the SRT task which requires continual response to all the stimuli presented. In this task, each sequence of motor responses (pressing a sequence of buttons) corresponds to a sequence of lights presented in different spatial locations. For the second task, we modified the original task so that it did not require the motor performance of the sequence. This task requires a selective response to a particular stimulus. Here the sequence of lights presented does not require a continual response to the sequence, but just to a particular location. Therefore, in the standard motor version of the SRT (SRTm) task, the sequence is learned by continuous motor reproduction. The reduced reaction time to all components of the sequence indicates the learning of the sequence. However, in the non-motor version of the SRT task (SRTnm), the sequence is not reproduced motorically but is only presented repeatedly, and thus reduced reaction time to a particular location indicates learning of the sequence. These tasks will be described in more detail in section 2.

In all the studies that tested the basal ganglia hypothesis of procedural memory in humans, participants were either PD or HD patients. The problem with testing these types of patient groups is that several studies have demonstrated that the pathology in PD [22] and HD [21] patients may extend beyond the basal ganglia region. For this reason, in this study we tested patients with cerebrovascular accident circumscribed to the basal ganglia. The goal of this study is twofold: First, to test patients with lesions restricted to the area of the basal ganglia and second, to test motor and non-motor versions of the same procedural learning task (i.e., SRTm and SRTnm) in addition to declarative memory tests.