Impact of Huntington's Disease on Mental Rotation Performance in Motor Pre-Symptomatic Individuals

Abstract

Background: Huntington's disease (HD) is a genetic disorder known for affecting motor control. Despite evidence for the impact of HD on visual cortico-striatal loops, evidence for impaired visual perception in early symptomatic HD patients is limited; much less is known about what happens during the HD prodrome.

Objective: The goals of this study were to evaluate perceptual processing in motor pre-manifest HD gene-carriers (Pre-HDs) during a visual mental rotation task.

Methods: To achieve this goal, 79 participants including 24 Pre-HD participants and 55 healthy matched controls were scanned using functional MRI as they performed a mental rotation task. Another group of 36 subjects including 15 pre-HDs and 21 healthy age/gender matched controls participated in a control behavioral test of judgment of line orientation outside the scanner.

Results: We found that, although Pre-HDs (in this stage of disease) did not demonstrate slower response times, their response accuracy was lower than controls. On the fMRI task, controls showed a significant decrease in activity in the occipito-temporal (OT) visual network and increase in activity in the caudo-fronto-parietal (CFP) network with mental rotation load. Interestingly, the amount of mental rotation-related activity decrease in the OT network was reduced in Pre-HDs compared to controls while, the level of CFP response remained unchanged between the two groups. Comparing the link between the evoked BOLD activity within these networks and response accuracy (i.e., behavior), we found that the models fit to data from controls were less accurate in predicting response accuracy of Pre-HDs.

Conclusion: These findings provide some of the earliest functional evidence of impaired visual processing and altered neural processing underlying visual perceptual decision making during the HD prodrome.

Keywords: Mental rotation; Pre-HD; functional MRI; visual system.

Fig. 1.

Experimental procedure and participants’ response accuracy during the mental rotation task. A) Stimuli were 10 different line drawing shapes subtended 4° × 4° on the screen. B) In each trial, two stimuli were presented side-by-side with 4° center-to-center distance. In 50% of trials they were matched, and in the rest of trials they were non-matched. C) Response accuracy of Pre-HDs (green) and healthy matched controls (red) decreased with increase in the mental rotation degree, and Pre-HD showed a lower response accuracy relative to healthy matched controls. D) In all mental rotation conditions, response accuracy of Pre-HD individuals decreased with increase in disease burden score [34]. Error bars show one standard error of mean.

Fig. 2.

Activity evoked during those trials in which mental rotation was not required (i.e., 0° mental rotation) in healthy controls (left columns) and Pre-HD individuals (right columns). In Red-to-yellow colors represent p < 10⁻³ to p < 10⁻⁶ in controls and p < 0.05 to p < 10⁻³ in Pre-HDs. This threshold difference is due to the higher number of controls compared to Pre-HDs. Otherwise, in both groups, we found a strong activity within occipital, occipito-temporal, parietal and frontal regions without any statistically significant (p < 0.05) difference between the two groups. Lack of activity within the foveal region (e.g., occipital pole) is due to peripheral stimulus presentation in our paradigm (see Methods). In each column, lateral, medial and flattened cortical views are presented in top, middle and bottom rows, respectively.

Fig. 3.

Brain areas that showed a significant response to mental rotation in healthy matched controls. Top) Activity map (p-values) evoked by contrasting the evoked response during trials in which stimuli were rotated 80° relative to each other, compared to trials in which stimuli were presented from the same angle (i.e., 0° mental rotation). Blue-to-cyan and red-to-yellow colors indicate those regions that show respectively an increase and a decrease in their level of FMRI activity during mental rotation relative no rotation trials. Activity map is overlaid on the inflated right hemisphere of the average brain (see Methods). Similar activity map was also evoked within the left hemisphere (not shown here). Bottom) Mental rotation related activity maps generated by contrasting the activity evoked during 80°, 60°, 40° and 20° mental rotation trials relative to activity evoked during 0° mental rotation trials in controls, overlaid on flattened cortex (right hemisphere). All maps are corrected for multiple comparisons. 1: cuneus and superior occipital gyri, 2: lingual gyrus and posterior transverse collateral sulcus, 3: fusiform gyrus, 4: orbitofrontal cortex, 5: superior frontal cortex, 6: superior parietal cortex, 7: inferior parietal cortex, 8: fronto-central cortex, 9: mid-anterior cingulate cortex, 10: anterior insular cortex.

Fig. 4.

Between-groups (Pre-HDs - Controls) comparison of activity evoked by mental rotation. A) Mental rotation (80° – 0°) evoked a larger (more negative) response in healthy controls relative to Pre-HD individuals within early visual areas, centered anteriorly relative to borders of the primary visual area (V1; [50]). The activity map (p-values) is corrected for multiple comparison. B) Border of regions of interest, defined based on subjects’ own high-resolution structural scans (i.e., defined independently from the functional responses), overlaid on an average inflated brain. Red and cyan outlines indicate ROIs that showed a higher and lower activity during mental rotation compared to not rotated trials, respectively. C) BOLD activity measured within cortical regions-of-interest (ROIs) of Pre-HDs (green) and healthy matched controls (red) across different mental rotation conditions relative to 0° trials. Consistent with the activity maps, impaired mental rotation related activity was found within early visual areas. However, this sensitive test also showed impaired activity within superior frontal cortex and orbitofrontal cortex of Pre-HDs relative to controls (Table 2). Error bars show one standard error of mean.

Fig. 5.

Mental rotation related activity measured within attention control area (superior and inferior parietal, fronto-central and mid-anterior cingulated cortex) and striatal nuclei (caudate, putamen, pallidum and accumbens) of controls (red) and Pre-HDs (green). Activity is measured during mental rotation trials relative to 0° trials. In this stage of disease, the between-groups difference remained non-significant in all of these brain regions (see also Table 2).

Fig. 6.

Relationship between the measured disease burden score and evoked brain activity in Pre-HDs across different mental rotation conditions. Among ROIs (i.e., those brain regions that showed a significant effect of rotation (Fig. 3)), activity within fronto-central and caudate, but not early visual areas, showed a significant correlation with disease burden score in all mental rotation degrees (see also Table 3). In these regions, the level of BOLD activity decreased as disease burden score increased. In each panel, each dot represents BOLD activity measured in the target ROI of one individual Pre-HD subject.

Fig. 7.

Impact of HD on cortical thickness and size of subcortical nuclei of Pre-HD individuals compared to healthy matched controls. A) The significance of cortical thinning in Pre-HDs compared to controls in the left and right hemispheres. We did not find any cortical thinning within early visual regions that showed functional impairments in Pre-HDs vs. controls outlined by black lines. Similar results were also obtained in the left hemisphere (not shown here). B) We also did not find any decrease in size of subcortical nuclei in Pre-HD (green) individuals compared to healthy controls (red). Notably, all measurements were normalized relative to the individual subject’s brain volume. These results are consistent with the fact that gene-expanded individuals participated in our tests were all in early motor pre-symptomatic stages of HD.

Fig. 8.

Results of predicting subjects’ behavior, based on the level of evoked BOLD response. We used a linear regression model to predict subjects’ response accuracy based on BOLD activity measured within the ROIs. Model was optimized based on the brain activity and response accuracy of 40 randomly selected healthy controls, and tested against the rest of controls and 15 randomly selected Pre-HDs. In all repetitions (n = 1000), model showed a better performance predicting the response accuracy (mean square error) of healthy controls (red) rather than Pre-HDs (green). Error bars show one standard error of mean.