Multimodal fusion technologies offer potential for detecting early micro-injuries in stroke and traumatic brain injuryNew Fusion Technologies May Detect Early Brain Injuries Better

Early-stage brain injury encompasses subtypes such as mild traumatic brain injury (mTBI), silent stroke and subclinical brain injury. It is characterised by negative imaging findings, subtle symptoms and a high risk of long-term neurological deficits, presenting a key challenge in the early screening and prevention of neurological diseases. Traditional imaging techniques, such as structural Magnetic Resonance Imaging (MRI) and Computed Tomography (CT), can only identify organic morphological lesions. Their detection rate for early micro-injuries at the level of neuronal electrophysiology and brain network function is extremely low. Consequently, these techniques struggle to meet the clinical needs for early screening, assessment, and intervention. Electroencephalography (EEG) spectral changes (including conventional spectral slowing and 1/f aperiodic abnormalities) and large-scale dynamic neural network disruption across the entire brain are core pathological features of early brain functional damage; they may serve as indirect correlates of microscopic pathological changes such as an imbalance in neuronal excitation-inhibition and abnormalities in the efficiency of brain information transmission. Based on this, this paper proposes a Dynamic-Spectral Coupling (DSC) framework, which proposes spectral abnormalities as a potential upstream mechanism associated with network reorganisation. Within this framework, this narrative review summarizes dynamic functional magnetic resonance imaging (dynamic fMRI), EEG spectral analysis and multimodal fusion technologies, with a particular focus on elucidating the coupling mechanisms between EEG periodic and 1/f aperiodic spectral parameters and dynamic brain network indicators. The review outlines the screening value and application advantages of multimodal fusion technologies in mTBI, occult cerebrovascular injury and subclinical brain injury. Furthermore, the paper summarises the current limitations of these technologies in terms of clinical translation, methodology and database development. Taking into account the clinical needs of neurorehabilitation, it proposes future development directions, including the development of lightweight devices, AI-assisted diagnosis, the establishment of standardised analytical systems and the translation of precision rehabilitation. The aim is to provide technical guidance and theoretical support for screening, risk stratification and rehabilitation planning in early-stage brain injury, while acknowledging that clinical utility remains to be established.