Implicit learning impairment identified via predictive saccades in Huntington's disease correlates with extended cortico-striatal atrophy.

The ability to anticipate events and execute motor commands prior to a sensory event is an essential capability for human's everyday life. This implicitly learned anticipatory behavior depends on the past performance of repeated sensorimotor interactions timed with external cues. This kind of predictive behavior has been shown to be compromised in neurological disorders such as Huntington disease (HD), in which neural atrophy includes key cortical and basal ganglia regions. To investigate the neural basis of the anticipatory behavioral deficits in HD we used a predictive-saccade paradigm that requires predictive control to generate saccades in a metronomic temporal pattern. This is ideal because the integrity of the oculomotor network that includes the striatum and prefrontal, parietal, occipital and temporal cortices can be analyzed using structural MRI. Our results showed that the HD patients had severe predictive saccade deficits (i.e., an inability to reduce saccade reaction time in predictive condition), which are accentuated in patients with more severe motor deterioration. Structural imaging analyses revealed that these anticipatory deficits correlated with grey-matter atrophy in frontal, parietal-occipital and striatal regions. These findings indicate that the predictive saccade control deficits in HD are related to an extended cortico-striatal atrophy. This suggests that eye movement measurement could be a reliable marker of the progression of cognitive deficits in HD.