Impaired  executive  organization  of  visual  search  contributes  to  abnormal  cognitive  motor  control   Tarkeshwar Singh, Christopher Perry, Sarah Tryon, Angela Ross and Troy Herter Department of Exercise Science, University of South Carolina, Columbia, SC-29208  

Frontoparietal networks mediate cognitive motor control, i.e. the processing and integration of perceptual, cognitive, and motor information to plan and execute goal-directed motor behavior1,2. Because of their broad distribution and interconnectivity, even focal damage to frontoparietal networks can cause simultaneous deficits in multiple cognitive and sensorimotor processes. Recent developments in theoretical motor control3 combined with connectivity-based approaches4 offer a novel framework to augment our understanding of the interaction between cognitive and motor processes during goal-directed movements. The purpose of this study was to investigate how stroke-induced impairments in executive organization of visual search contribute to deficits in cognitive motor control. We studied visuomotor performance of 69 participants, including 37 young controls (21-50 yrs), 16 older controls (51-80 yrs) and 16 stroke survivors (48-80 yrs) who had frontoparietal lesions caused by an MCA stroke. Participants performed the Trail-Making Test5 (TMT) on a KINARM End-Point robot with integrated eye tracking (Fig. 1). TMT is a visuomotor task that consists of two parts: a numeric variant (TMT-A) in which participants make sequential reaching movements between the first 25 positive natural numbers (1, 2,...25), and an alphanumeric variant (TMT-B) in which they alternatingly connect the first 13 natural numbers and first 12 Roman letters (1, A, 2, B,…13). Subjects are instructed to complete each variant in the shortest possible time. Quick completion of the TMT requires efficient visual search, quick decision-making, and rapid reaching movements. These features of TMT also make it a strong predictor of success at on-road driving tests following stroke6. To minimize confounds due to motor impairments, stroke survivors used their less affected hand (controls used dominant hand). We measured the total time (Total Time) to complete the test. Total Time was further divided into dwell time, i.e. the time the hand stayed on a number/letter and movement time, which is the time spent reaching between them. From the eye-tracking data, we measured the Number of Saccades and Fixation Duration at each number/letter. Fig. 2 shows a stem plot of the eye and hand movements for three exemplar participants. This Figure shows that the young control displayed the shortest Total Time and least Number of Saccades, whereas the stroke survivor exhibited the longest Total Time and greatest Number of Saccades. This relationship between Total Time and Number of Saccades suggests that abnormal performance (long Total Time) depends on deficits in visual search (greater Number of Saccades). This is supported by Fig. 3, which shows that on average, stroke survivors exhibited longer Total Time (Fig. 3A) and a greater Number of Saccades (Fig. 3B) than controls. Furthermore, stroke survivors displayed longer fixation durations (Fig. 3C), suggesting longer times for decision-making than controls. Between the two test variants, stroke survivors made more saccades and slower reaching movements during TMT-B. The slow movements were accompanied by more saccades during Movement Time and jerkier hand movements (many velocity peaks/movement). This suggests that during increased cognitive load in TMT-B, visual search interfered with hand movements in stroke survivors. To investigate how frontoparietal lesions influenced executive organization of visual search, we created a stochastic model that predicted how participants used topographic planning strategies and working memory to organize their visual search. The output of the model were two parameters: WM (Working Memory: size of the working memory buffer) and TS (Topographic Search Strategies: size of the visual search span used to search for targets). Fig. 4 shows that most stroke survivors did not use working memory and searched for targets over a larger area. This indicates that the main determinant of abnormal TMT performance in stroke survivors was their inability to combine working memory (WM) with topographic strategies (TS) during visual search. Accordingly, TS and Number of Saccades were the strongest predictors of TMT performance (Total Time). Finally, the average Fixation Duration, a measure of sensorimotor decision-making, was not correlated with Total Time suggesting that slow decision times had a minimal influence on task performance. To underscore the importance of our results for visuomotor coordination, we discuss our findings within two conceptual frameworks, the serial information processing7,8 approach and the Affordance Competition Hypothesis3. References: 1. Siegel, M. et al. (2012). Nat Rev Neurosci, 13(2), 121-34. 2. Georgopoulos, A. P. (2000). Curr Opin Neurobiol, 10, 238-41. 3. Cisek, P. & Kalaska, J. F. (2010). Annu Rev Neurosci, 33, 269-98. 4. Carter, A. R et al. (2012). NeuroImage, 62, 2271-80.

5. Reitan, R. M. (1958). Percept Motor Skill, 8, 271-6. 6. Devos, H. et al. (2011). Neurology, 76(8), 747-756 7. Land, M. F. (2009). Visual Neurosci, 26, 51-62. 8. Song, J.-H., et al. (2009). Trends Cogn Sci, 13, 360-366.

 

Figure 1: The setup of the Trail-Making Test. TMT-A is shown on the left panel and TMT-B on the right.

Figure 3: This Figure shows the descriptive statistics of the overall measure of performance (Total Time, panel-A) and two measures from the eye-tracker. Each circle represents a participant. The red line is the mean and cyan and yellow areas indicate 1 and 2 SD respectively. Panel B shows the total number of saccades made by the participants. Panel C shows the distribution of fixation durations for the three groups.

Figure 2: A stem plot representation of the eye and hand movement for TMT-A for three representative participants from each group. Xaxis denotes the targets (1,2…25) and Y-axis denotes time. Green circles indicate fixations on targets and blue and gold squares indicate hand dwell on target. Large radii of circle(s) denote longer fixation durations. Deficits in visual search (Number of Saccades), sensorimotor decision-making (circle radii) and overall performance (much longer Total Time) are evident in the stroke survivor.

Figure 4: The participants were divided into three groups based on cluster analysis. Here ‘TS’ and ‘WM’ are plotted against the log (Total Time). The green cluster consisted mostly of controls and only three stroke survivors. This group used working memory (WM≥1) but performed the visual search in a larger area than the red cluster. The red cluster also comprised mostly of controls and this group did not use working memory (WM=0) but performed the visual search in a much smaller area (small TS). The last cluster (blue) consisted mostly of stroke survivors and participants in this cluster did not use working memory (WM=0) but performed visual search in a much larger area (large TS) than the other two clusters. Also note that TS and log(Total Time) are strongly correlated (ρ=0.85,p<0.001). This Figure shows that the inability to combine working memory with visual search area contributed to poor performance on the TMT.