Distinct Patterns of Cortical Reorganisation Accompany Behaviour-Specific Development in Freely Behaving Infants

A central premise of developmental neuroscience is that mature behaviour emerges through reciprocal interactions between infants spontaneous actions, neural activity, and environmental experience. Yet, much human neuroimaging research relies on highly controlled, repetitive paradigms designed to minimize behavioural variability, obscuring the very brain-behaviour dynamics that drive developmental change. Here, we introduce a naturalistic functional near-infrared spectroscopy (fNIRS) paradigm that exploits, rather than suppresses, natural variability in infants behaviour to investigate cortical changes associated with behavioural development. Continuous video coding of spontaneous infant behaviour was combined with finite impulse response (FIR) modelling to estimate behaviour-specific hemodynamic responses associated with reach-to-grasp, hand-to-mouth, and vocalization behaviours while accounting for temporally overlapping hemodynamic responses by treating extraneous behaviours as nuisance regressors. From 6 to 13 months of age, reach-to-grasp and vocalizations increased in frequency, while all three behaviours shortened in duration. Critically, developmental changes in cortical activation were behaviour-specific. During reach-to-grasp movements, bilateral posterior parietal activity decreased with age. In contrast, during hand-to-mouth and vocalizations, somatomotor and premotor activity increased with age. Cortical activity was also related to individual variation in the frequency and duration of these behaviours, indicating that cortical recruitment remained coupled to behavioural variation independent of age. These findings reveal behaviour-specific patterns of cortical reorganisation during infant development and establish a methodological framework for linking spontaneous behaviour to cortical activity in freely behaving infants. More broadly, this approach enables direct investigation of cortical development as it unfolds through naturalistic interactions among behaviour, neural activity, and experience. Significance StatementContinuous self-generated interactions with the environment play a critical role in driving infant brain development, yet neuroimaging research often examines infants while sleeping or in highly constrained and repetitive tasks. Here, we show that spontaneous infant behaviour can instead be leveraged to provide the temporal variability needed to resolve behaviour-specific cortical responses in freely behaving infants. By combining continuous behavioural coding with data-driven finite impulse response (FIR) modelling of overlapping hemodynamic responses, we identified behaviour-specific developmental differences in cortical recruitment without tightly constraining the infants behaviour. This novel approach expands developmental neuroimaging beyond sleeping or trial-based paradigms and demonstrates how naturally occurring behavioural variability can become a useful feature of fNIRS experimental design. More broadly, it generates new opportunities to directly test theories of neurobehavioural development in more ecologically valid settings.