Dorsal visual stream is preferentially engaged during externally guided action selection in Parkinson Disease

Abstract

Objective: In patients with Parkinson Disease (PD), self-initiated or internally cued (IC) actions are thought to be compromised by the disease process, as exemplified by impairments in action initiation. In contrast, externally-cued (EC) actions which are made in response to sensory prompts can restore a remarkable degree of movement capability in PD, particularly alleviating freezing-of-gait. This study investigates the electrophysiological underpinnings of movement facilitation in PD through visuospatial cuing, with particular attention to the dynamics within the posterior parietal cortex (PPC) and lateral premotor cortex (LPMC) axis of the dorsal visual stream.

Methods: Invasive cortical recordings over the PPC and LPMC were obtained during deep brain stimulation lead implantation surgery. Thirteen PD subjects performed an action selection task, which was constituted by left or right joystick movement with directional visual cuing in the EC condition and internally generated direction selection in the IC condition. Time-resolved neural activities within and between the PPC and LPMC were compared between EC and IC conditions.

Results: Reaction times (RT) were significantly faster in the EC condition relative to the IC condition (paired t-test, p = 0.0015). PPC-LPMC inter-site phase synchrony within the β-band (13-35 Hz) was significantly greater in the EC relative to the IC condition. Greater PPC-LPMC β debiased phase lag index (dwPLI) prior to movement onset was correlated with faster reaction times only in the EC condition. Multivariate granger causality (GC) was greater in the EC condition relative to the IC condition, prior to and during movement.

Conclusion: Relative to IC actions, we report relative increase in inter-site phase synchrony and directional PPC to LPMC connectivity in the β-band during preparation and execution of EC actions. Furthermore, increased strength of connectivity is predictive of faster RT, which are pathologically slow in PD patients. Stronger engagement of the PPC-LPMC cortical network by an EC specifically through the channel of β-modulation is implicated in correcting the pathological slowing of action initiation seen in Parkinson's patients.

Significance: These findings shed light on the electrophysiological mechanisms that underlie motor facilitation in PD patients through visuospatial cuing.

Keywords: Action selection; Choice task; Dorsal visual stream; Electrocorticography; Posterior parietal cortex; Premotor cortex.

Figure 1.

Schematic of task design. 1) Black screen for duration of 1.0 to 1.1 seconds (s) during which patient is resting. 2) Fixation cross for duration of 1.0 to 1.1s indicates that cue to move is impending. 3) Onset of action selection cue indicates patient should move joystick. In the internally cued (IC) condition bidirectional arrows are displayed indicating that patient may move joystick left or right. In the externally cued condition (EC) uniform right or left pointing arrows are displayed, instructing action selection. Cue is displayed for 1.5s or until response is recorded. 4) Response is recorded for additional 0.5s during display of black screen. In total, response is recorded for 2s following action selection cue display. After the completion of a trial, a new trial begins.

Figure 2.

Group averaged projection of posterior parietal cortex (A, B) and lateral premotor cortex (C,D) electrode recording area over standard Montreal Neurological Institute standard brain in lateral (A,C) and dorsal (B,D) views. Red points demonstrate the position of electrodes. Central sulcus is marked with red line.

Figure 3.

Cue locked power in the posterior parietal cortex (PPC) and lateral premotor cortex (LPMC). Movement-related beta suppression was observed in both PPC and LPMC. This suppression was not cluster significant in the PPC in the EC condition. In the PPC, there were no significant power differences between externally cued (EC) and internally cued (IC) conditions. In the LPMC there was significantly greater beta suppression toward the end of the trial. Black contour indicates cluster significance. X-axis is time from cue. Y-axis is frequency. Color indicates group averaged z-power. Solid vertical line indicates action selection cue onset, first dotted vertical line indicates movement onset, reaction time (RT). Dashed vertical lines indicates maximal joystick displacement and movement end. Hz, hertz; Mov, movement; s, seconds.

Figure 4.

A) Cue locked beta-band debiased weight phase lag index (dwPLI) between posterior parietal cortex (PPC) and lateral premotor cortex (LPMC). Significant differences were observed between externally cued (EC) and internally cued (IC) conditions, occurring around the time of movement onset, and sustained until peak joystick displacement. X-axis is time from cue. Y axis is z-dwPLI. Time periods with significant dwPLI differences between conditions are highlighted with green. Solid vertical line indicates timing of action selection cue. Bold dotted vertical line indicates group averaged movement onset, reaction time (RT). Gray dashed vertical lines indicate maximal joystick displacement and movement end. B) Spearman R values over time. Y axis is cross-subject spearman R value correlating beta-band PPC-LPMC dwPLI at each time point versus RT. Significant R values in the EC condition are highlighted in blue. No significant R values were observed in the IC condition. Significant differences in R values across EC and IC conditions are highlighted in green. C) Cue locked Beta-band Granger Causality (GC) in PPC and LPMC. Y-axis is z-GC. Significant differences between EC and IC conditions are highlighted in green. Max, maximum; Mo, movement; S, seconds.