Early prediction of long-term tactile object recognition performance after sensorimotor stroke

Area of Science:

• Neuroscience
• Neurology
• Radiology

Background:

• Tactile agnosia, the inability to recognize objects by touch, has been primarily studied through small case studies.
• Cortical sensori-motor stroke can lead to significant hand skill impairment and limitations in daily activities, affecting tactile object recognition (TOR).

Purpose of the Study:

• To investigate the utility of multi-voxel pattern analysis (MVPA) on early diffusion-weighted lesion maps for predicting long-term TOR recovery in stroke patients.
• To identify specific dysfunctional subprocesses and critical brain regions involved in impaired TOR after stroke.

Main Methods:

• Applied MVPA to diffusion-weighted lesion maps from 35 stroke subjects to predict TOR performance.
• Utilized multiple regression analysis to identify dysfunctional subprocesses (grasping, macrogeometrical perception).
• Combined MVPA with a priori knowledge of regions of interest for shape recognition to define functionally relevant voxels.

Main Results:

• MVPA accurately predicted TOR recovery trajectories over nine months, identifying three patient recovery subgroups.
• Baseline motor hand skill and shape discrimination significantly correlated with TOR recovery.
• Lesion sites in persistently impaired subjects overlapped with critical areas (PFt, OP1) for higher-order sensory processing.

Conclusions:

• Early neuroimaging data, specifically MVPA of lesion maps, can reliably forecast post-stroke TOR recovery.
• This predictive model has the potential to generalize to larger patient cohorts with first cortical ischemic stroke.
• The findings highlight the clinical utility of neuroimaging for individual recovery forecasting and personalized rehabilitation planning.