Abstract Background Hand dexterity impairment is a key feature of disability in people with primary progressive multiple sclerosis (PPMS). So far, ocrelizumab, a recombinant humanized monoclonal antibody that selectively depletes CD20-expressing B cells, is the only therapy approved for PPMS and recent analysis reported its ability to reduce the risk of upper limb disability progression. However, the neural mechanisms underlying hand impairment in PPMS and the brain networks behind the effect of ocrelizumab on manual dexterity are not fully understood. Objective Main aims of our study were: (i) to investigate neurophysiological and behavioural correlates of hand function impairment in subjects with PPMS, and (ii) to use neurophysiologic and behavioural measures to track the effects of ocrelizumab therapy on manual dexterity. Methods Seventeen PPMS patients and 17 healthy-controls underwent routine neurophysiological protocols assessing the integrity of cortico-spinal and somatosensory pathways and advanced transcranial magnetic stimulation (TMS) protocols evaluating inhibitory (short and long interval intracortical inhibition, short-latency afferent inhibition) and facilitatory (motor thresholds, intracortical facilitation, short-interval intracortical facilitation) circuits in the primary motor cortex. All subjects also underwent behavioural analysis of hand dexterity by means of nine-hole peg test and finger movement analysis, and hand strength with handgrip and three-point pinch test. Neurophysiological and clinical assessments of hand functionality were also performed after 1 year of ocrelizumab therapy. Results At baseline PPMS patients displayed a significant impairment of hand dexterity and strength compared to healthy controls (all p < 0.03). Neurophysiological study disclosed prolonged latencies of standard somatosensory and motor evoked potentials (all p < 0.025) and an overall reduction of intracortical excitability at TMS protocols, involving both excitatory and inhibitory circuits. Importantly, hand dexterity impairment, indexed by delayed 9HPT, correlated with TMS protocols investigating cortical sensorimotor integration (short-latency afferent inhibition, SAI), p = 0.009. Both parameters, 9HPT (p = 0.01) and SAI (p = 0.01), displayed a significant improvement after 1 year of therapy with ocrelizumab. Conclusion Intracortical sensorimotor networks are involved in hand dexterity dysfunction of PPMS. Ocrelizumab therapy displays a beneficial effect on hand dexterity impairment most likely through intracortical networks implicated in fast sensorimotor integration.