Divergent Rehabilitation Targets in Pediatric Cerebral Palsy: Asymmetry-Bounded Versus Symmetry-Disrupting Strategies in Hemiplegic and Diplegic Gait

Background/Objective: Rehabilitation for gait asymmetry in cerebral palsy (CP) commonly treats improvements in gait symmetry indices as a subtype-general marker of therapeutic success. Building on a trajectory-divergence model developed in a companion review, we propose that hemiplegic and diplegic CP represent opposing developmental deviations from the normative gait-symmetry trajectory — persistent, actively regulated compensatory asymmetry in hemiplegia versus pathological rigid symmetry in diplegia — such that a single, shared correction target is conceptually misaligned for at least one subtype. This review critically evaluates, through a narrative synthesis of literature published between 2010 and 2026, whether rehabilitation strategies for hemiplegic and diplegic CP should be differentiated according to this model rather than converging on gait-symmetry restoration. Methods: A structured narrative review following a Population–Concept–Context (PCC) framework and JBI-aligned appraisal, drawing on 70 unique sources across 21 thematic searches. Results: In hemiplegic CP, compensatory asymmetry is mechanistically and energetically dissociable from correction, but its magnitude, not its presence, predicts balance and locomotor outcomes, bounding rather than eliminating the case for conservation. In diplegic CP, restoring adaptive motor variability is supported by convergent evidence but is conditional on confirming genuine bilateral rigidity, since a meaningful subset of diagnosed diplegic children presents with clinically significant inter-limb asymmetry. A provisional Divergent Rehabilitation Response Index (DRRI) organizes four domains of subtype-specific therapeutic congruence: counter-evidence, such as correction-beneficial cases, inconsistent functional transfer, evidentiary concentration, and long-term musculoskeletal risk signals, are directly addressed. Conclusions: Differentiated magnitude- and subtype-conditional rehabilitation targets, rather than universal symmetrization, are proposed for hemiplegic and diplegic CP, alongside an explicit research agenda.