The ability to adapt behavior to dynamic environments necessitates the flexible adjustment of both decision and movement speeds. While prior research suggests a significant role for the basal ganglia, and particularly the subthalamic nucleus (STN), in modulating these processes, the extent to which these are controlled by common or distinct mechanisms remains unclear. The STN receives afferents from various cortical areas involved in decision-making and motor control. A prevailing hypothesis proposes that the STN contributes to response selection and decision speed by withholding a motor response until sufficient evidence is gathered or response conflicts are resolved. Separately, research indicates a key role for the basal ganglia in modulating movement vigor, specifically speed. Although these functions share commonalities in determining the time to reach a goal, they are often studied independently. This study aims to investigate whether STN control of decision and movement speed involves distinct or shared mechanisms.

Existing literature points towards the basal ganglia's involvement in both decision-making and motor control. Studies using computational models and empirical data suggest the STN's crucial role in determining when to terminate deliberation and the vigor of movement execution. However, the precise neural mechanisms involved in adjusting decision and movement speeds, and whether common or separate signals are used, remain unresolved. Previous correlative studies have explored STN activity and its relationship to decision and movement speed, with some suggesting temporal precedence of STN involvement in decision speed over movement speed. These studies, however, are largely correlative in nature and lack direct causal evidence. This study aims to provide further clarity on these questions.

This study employed a perceptual decision-making task where participants indicated the direction of moving dots by pressing a dynamometer. Thirteen Parkinson's disease (PD) patients implanted with deep brain stimulation (DBS) electrodes and fifteen age-matched healthy controls (HC) participated. The task included instructions to prioritize speed or accuracy. Force recordings allowed the separation of response times into reaction time (cue to movement onset) and movement time (movement onset to peak force). In PD patients, STN local field potentials (LFPs) were recorded, and bursts of electrical STN stimulation were applied in separate sessions. A hierarchical drift-diffusion model (HDDM) was used to analyze the behavioral data and to understand the underlying decision-making parameters. STN LFP data were analyzed in the frequency domain to identify relationships between beta, theta, and gamma power and behavioral adjustments. Cluster-based permutation tests were used to analyze the effects of stimulation on reaction and movement time adjustments across different time windows. Finally, unilateral stimulation was applied to determine hemispheric specificity of STN control.

The study reveals that both PD patients and HC participants adjusted their decision and movement speeds according to instructions, with speed instructions leading to significantly shorter reaction and movement times. HDDM analysis showed that speed instructions reduced decision thresholds, allowing responses at lower levels of evidence. A significant negative correlation was found between decision thresholds and movement times, suggesting that higher certainty choices were accompanied by faster movements. STN beta activity was related to both reaction and movement time adjustments, but in distinct temporal windows. Cue-aligned beta power decrease was strongly predictive of shorter reaction times and lower decision thresholds. Movement-aligned beta power predicted faster movements, particularly under speed instructions. STN burst stimulation causally affected reaction and movement time adjustments but only in specific time windows; DBS applied before the cue impaired reaction time adjustments, while stimulation applied after the cue affected movement time adjustments. Unilateral stimulation demonstrated that STN predominantly controls decision speed for the contralateral hemisphere, similar to its control over movement. These stimulation effects were accompanied by specific alterations in STN beta power.

The findings demonstrate that the STN independently controls decision and movement speed through distinct temporal windows, supporting the hypothesis of separate mechanisms. The strong lateralization of STN effects on decision speed aligns with the known lateralized control of movement, suggesting a hierarchical model where cortical computations set a global decision threshold while the STN fine-tunes these thresholds at a hemispheric level. The negative correlation between decision thresholds and movement time may reflect factors beyond STN activity, such as better movement preparation during longer reaction times. The similar behavioral adjustments in both PD patients and HC participants suggest that these findings might generalize beyond the patient population studied. The independent control of decision and movement speeds allows for flexible behavioral adaptation to various demands.

This study provides compelling evidence for the independent control of decision and movement speed by the STN within distinct temporal windows. The findings highlight the STN's role in flexible behavioral control and its lateralized influence on decision speed. Future research could investigate the specific neural circuits involved in the STN's control over decision-making and movement and explore the potential for targeted neuromodulation strategies to improve motor and cognitive performance.

The study's limitations include the use of burst stimulation, which might not have allowed STN neurons to reach a stable baseline. The study also used clinically effective stimulation intensities, potentially overlooking effects at lower intensities. Finally, the task used might evoke stronger changes in reaction times than in accuracy rates. Future research should address these points by using more precise stimulation paradigms and tasks.