Parkinson's disease is characterized by progressive degeneration of dopaminergic neurons within the nigrostriatal pathway, leading to motor and non-motor deficits that become less responsive to chronic dopaminergic pharmacotherapy. Here, we examined whether a multimodal mobile digital intervention, DopApp™, delivering adaptive daily training across sensorimotor, psychological and rehabilitation domains, can induce functional reorganization of thalamocortical motor and limbic networks and augment dopaminergic treatment effects in Parkinson's disease. To test this hypothesis, we conducted a pilot study in 42 levodopa-treated patients randomized to a 3-week DopApp™ or placebo app intervention. Clinical, psychological and resting-state functional MRI assessments were acquired at baseline and post-treatment. Compared with the placebo group, the active group showed greater motor improvements, as assessed by total scores on the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS; P < 0.001; d = 1.22). These improvements were driven by gains in motor function (Part III: P < 0.001; d = 1.16) and activities of daily living (Part II: P < 0.05; d = 0.68). Depressive symptoms also improved during the intervention (Beck's Depression Inventory II: P < 0.05; d = 0.76), representing a parallel non-motor gain. Seed-based resting-state analyses revealed increased group-by-time connectivity within two parallel thalamocortical pathways. In the motor pathway, connectivity strengthened between the bilateral ventral lateral thalamic nuclei and primary motor cortex (P FDR < 0.05), correlating with motor improvement (MDS-UPDRS Part III: r = -0.72; P < 0.05). In parallel, increased connectivity was observed within the limbic circuit linking the anteroventral thalamus with medial prefrontal, posterior cingulate and hippocampal regions (P FDR < 0.05). Connectivity increases within this pathway correlated with improved depressive symptoms (r = -0.65; P < 0.01) and self-perceived daily function (MDS-UPDRS Part II: r = -0.66; P < 0.01). Additionally, engagement with specific app modules predicted domain-specific gains, and connectivity changes correlated with these improvements, suggesting a dose-response relationship and parallel circuit-specific neuroplasticity. These findings provide preliminary evidence supporting the integration of scalable digital interventions into precision, drug-digital treatment paradigms for Parkinson's disease and related neurodegenerative disorders. Moreover, they represent a conceptual advance towards personalized care models that enhance drug efficacy and expand therapeutic impact beyond conventional clinical settings.