Spatial-orientation priming impedes rather than facilitates the spontaneous control of hand-retraction speeds in patients with Parkinson's disease

Abstract

Background: Often in Parkinson's disease (PD) motor-related problems overshadow latent non-motor deficits as it is difficult to dissociate one from the other with commonly used observational inventories. Here we ask if the variability patterns of hand speed and acceleration would be revealing of deficits in spatial-orientation related decisions as patients performed a familiar reach-to-grasp task. To this end we use spatial-orientation priming which normally facilitates motor-program selection and asked whether in PD spatial-orientation priming helps or hinders performance.

Methods: To dissociate spatial-orientation- and motor-related deficits participants performed two versions of the task. The biomechanical version (DEFAULT) required the same postural- and hand-paths as the orientation-priming version (primed-UP). Any differences in the patients here could not be due to motor issues as the tasks were biomechanically identical. The other priming version (primed-DOWN) however required additional spatial and postural processing. We assessed in all three cases both the forward segment deliberately aimed towards the spatial-target and the retracting segment, spontaneously bringing the hand to rest without an instructed goal.

Results and conclusions: We found that forward and retracting segments belonged in two different statistical classes according to the fluctuations of speed and acceleration maxima. Further inspection revealed conservation of the forward (voluntary) control of speed but in PD a discontinuity of this control emerged during the uninstructed retractions which was absent in NC. Two PD groups self-emerged: one group in which priming always affected the retractions and the other in which only the more challenging primed-DOWN condition was affected. These PD-groups self-formed according to the speed variability patterns, which systematically changed along a gradient that depended on the priming, thus dissociating motor from spatial-orientation issues. Priming did not facilitate the motor task in PD but it did reveal a breakdown in the spatial-orientation decision that was independent of the motor-postural path.

Figure 1. Priming Experiment to increase the cognitive load of a simple reach-to-grasp task.

(A) Rods rendered in three dimensions were presented on the computer screen at 1 of 5 possible locations (shown here at the center location) at the four corners and at the center of the monitor. Color indicated the target speed (red-slow and green-fast). Speed was also labeled at the center of the cylinder. The target orientation could be horizontal or vertical. Because of redundancy in the degrees of freedom at multiple joints of the arm, each one of these oriented cylinders affords more than one arm-hand orientation. Subjects were free to choose the final orientation in the DEFAULT condition. (B) The primed condition instructed the subjects to use a particular target orientation while matching the hand-held cylinder to the simulated cylinder on the screen. The subjects were instructed to pick the orientation as though they were going grab the cup and drink from it. This instruction evoked a precise arm-hand orientation that was generally different from the DEFAULT one chosen by the subject in the first block. The primed-UP case required the same orientation as the DEFAULT but the primed-DOWN case required mental rotation to align the hand to the cup as if “picking it up to drink from it”. This orientation cue evoked rotations at the arm joints and at the hand that were unambiguously different from the DEFAULT and primed-UP cases. (C) An example of a DEFAULT arm-hand orientation evoked by the cylinder oriented vertically and positioned at the center of the screen. (D) The same canonical orientation of the cylinder evokes a very different arm posture and a different hand orientation during primed-DOWN.

Figure 2. Similar biomechanical constraints for DEFAULT and primed-UP in vertical and horizontal cases contrast with different biomechanical demands between primed-UP and primed-DOWN.

Traces are the averaged projection of the degrees of freedom of the arm along the dimensions relevant to the task goals and the dimensions incidental to the goals. These averages are taken across 100 trials for a typical representative and for 100 frames. Red continuous lines are task-relevant DoF forward. Blue are task-relevant DoF retractions and black are task-incidental DoF traces. Dashed lines are the standard deviation from the mean traces. During forward segments and retractions in the same loop the recruitment, release and balance of the degrees of freedom of the arm tend to markedly change as a function of task complexity.

Figure 3. Examples of sub-types of Parkinson’s disease as a function of site of Lewy Bodies (LB) appearance in sub-cortical and cortical structures, density of LB and time course of spread.

In all cases the presence of LB eventually leads to various stages of dementia and various subtypes of PD.

Figure 4. Individual stochastic signatures of variability for each patient separate the scatter in the Gamma plane representing the voluntary task-relevant (forward) and the spontaneous (retracting) task-incidental modes of acceleration control across all conditions.

The cognitive load of the task systematically alters the slope of the retracting segments with spatial orientation priming.

Figure 5. Effects of priming on the movement trajectories at the wrist in typical NC participant using two different levels of speeds randomly cued.

(A) DEFAULT forward motion trajectories with corresponding speed profiles for slow (red) and fast (green) cases. Inset shows the actual stimuli on the screen priming the subject to match the orientation of the rod on the screen (vertical in this case) with the hand-held rod. Top are the forward paths and bottom are the retracting motions. (B) Primed-UP cases evoked similar final orientations as the DEFAULT condition. The inset shows the priming cup with handle next to the original rod. This figure is made from the right hand data and right hand stimuli for a right-handed person. (C) The primed-DOWN condition changed the trajectories in both the forward and retracting cases. The inset shows the priming condition where the arm and hand underwent complex rotations. The instruction was to match the orientation of the rod on the screen as if the hand were to gasp the handle of the cup to drink from it. Notice the dramatic differences in trajectories for all target positions. NCs maintain the instructed speed throughout the continuous forward-and-back loop.

Figure 6. Effects of priming on the movement trajectories at the wrist in typical patient within PD2 group using two different levels of speeds randomly cued.

(A) DEFAULT forward motion trajectories with corresponding speed profiles for slow (red) and fast (green) cases. Inset shows the actual stimuli on the screen priming the subject to match the orientation of the rod on the screen (vertical in this case) with the hand-held rod. Top are the forward paths and bottom are the retracting motions. (B) Primed-UP cases evoked similar final orientations as the DEFAULT condition. The inset shows the priming cup with handle next to the original rod. (C) Primed-DOWN condition changed the trajectories in both the forward and retracting cases. The inset shows the priming condition where the arm and hand underwent complex rotations. The instruction was to match the orientation of the rod on the screen as if the hand were to gasp the handle of the cup to drink from it. Notice the dramatic differences in trajectories for all target positions. Retracting speed profiles in the primed-DOWN condition in PD2 group do not have statistically significant differences for instructed fast and slow speeds.

Figure 7. Effects of priming on the movement trajectories at the wrist in typical patient within PD1 group using two different levels of speeds randomly cued.

(A) DEFAULT forward motion trajectories with corresponding speed profiles for slow (red) and fast (green) cases. Inset shows the actual stimuli on the screen priming the subject to match the orientation of the rod on the screen (vertical in this case) with the hand-held rod. Top are the forward paths and bottom are the retracting motions. (B) Primed-UP cases evoked similar final orientations as the DEFAULT condition. The inset shows the priming cup with handle next to the original rod. (C) Primed-DOWN condition changed the trajectories in both the forward and retracting cases. The inset shows the priming condition where the arm and hand underwent complex rotations. The instruction was to match the orientation of the rod on the screen as if the hand were to gasp the handle of the cup to drink from it. Notice the dramatic differences in trajectories for all target positions. Retracting speed profiles in both primed-DOWN and primed-UP condition in PD1 group do not have statistically significant differences for instructed fast and slow speeds.

Figure 8. Self-emerging clusters and patient subtypes based on speed maxima (m/s).

(A) Representative NC and patients from PD1 and PD2 subgroups grouped the speed maxima differently in the forward and retracting motions as a function of cognitive load condition. The NC maintained consistent separation across conditions in both portions of the pointing gesture yet patients consistently performed worse in the primed cases even though primed-UP was biomechanically equivalent to DEFAULT. PD1 performed the worst with no distinction in the primed cases. Slopes changed systematically with cognitive loads even for biomechanically similar DEFAULT and primed-UP motions. (B) Self emerging subtypes of misclassified trials from the blind clustering k-means separated exactly as the p-value statistics had predicted (see Methods and Results for details).

Figure 9. Effects of increasing task difficulty on the speed and timing of the reach by priming the final desired orientation such as to evoke complex mental rotation of the stimulus.

(A) PD1 patients moved significantly slower than PD2 patients (median 0.65 m/s vs. 1.15 m/s, χ² 215.05, p<10^{− 48}) according to the values of the hand’s maximum speed returning from the target to the resting position. (B) The timing of the maximum speed was not significantly different in the two patient groups (median 0.52 s, χ² 2.86, p>0.1) but both groups took longer to reach the velocity peak than NC’s did (median 0.35 s, χ² 123.5, p<10^{− 27}).

Figure 10. Normalized frequency distribution of the maximum speed and maximum acceleration values from the retracting primed motions in two groups of patients with PD.

(A) Arm trajectories (shoulder, elbow, wrist, and hand) during the primed-UP condition similar to the DEFAULT case. The initial and final arm postures corresponding to the trajectories towards two randomly selected positions are superimposed. Patients in the PD1 group had a nearly symmetric distribution of speed maxima in the primed-UP condition where their retracting motions could no longer differentiate between the randomly instructed speeds. The group of PD2 was comprised of patients whose retracting motions could, on average, differentiate between instructed fast or slow speeds during the easier primed-UP cases. Their distribution of maximum speed values was skewed.

Figure 11. Individual stochastic signatures of variability for the retracting speed maxima in NC vs. patient types.

Notice the changes in slope and intercept with changes in the cognitive load induced by the priming. Primed UP separates patients in PD1 maximally from the NC and from most of the patients in PD2. Primed DOWN shifts the stochastic signatures of the speed maxima for all participants, (details of the linear fit in Table S3).