Identification of abnormal neural language networks by reading "brainprints" in patients with brain tumors

Alterations in neural language networks are common in patients with brain tumors, yet their nature varies substantially across individuals. By reducing data to group-level averages, conventional analyses fail to capture such heterogeneity, obscuring patient-specific information.

The present study applied a resting-state connectivity fingerprinting approach to characterize language network alterations at the single-subject level, yielding individualized connectivity profiles ("fingerprints"). Fingerprints of 27 right-handed patients with a left-hemisphere brain tumor affecting language-relevant areas were assessed at three time points (preoperative, immediate postoperative and three-month follow-up). Connectivity patterns were compared to a normative reference derived from 30 healthy participants and linked to language performance.

Fingerprints remained temporally stable in healthy individuals. In patients, fingerprints revealed distinct, patient-specific deviations from the typical network structure with highly heterogeneous changes over time. Three main findings emerged: (1) patients with language deficits showed greater deviations from the typical fingerprint than those without deficits; (2) significant associations between larger deviations and poorer language performance were confined to the immediate postoperative phase, likely reflecting surgery- or treatment-related influences or differences in the (mal)adaptivity of reorganization over time; (3) in high-grade glioma, exploratory analyses provided preliminary evidence for an adaptive contribution of the contralesional hemisphere immediately after surgery.

The findings support connectivity fingerprinting as a promising approach for characterizing patient-specific network patterns and monitoring functional reorganization processes relevant to language function at the single-subject level. With continued methodological refinement, this approach holds potential for contributing to more individualized clinical decision-making within the context of personalized medicine.