Shedding light on imagined scents: an fNIRS feasibility study of olfactory mental imagery and semantic cueing

Mental imagery is the ability to evoke sensory experiences without a corresponding stimulus, with established roles in memory, cognition, and behavior across multiple sensory modalities. Whilst visual and auditory imagery have been extensively characterized using neuroimaging, olfactory imagery remains comparatively understudied despite representing a particularly compelling target for investigation. Unlike other sensory systems, olfactory information projects directly to limbic structures including the amygdala and hippocampus without thalamic relay, providing a neurobiological basis for the characteristically vivid and emotionally evocative quality of odor-related memories and imagery. Understanding the cortical correlates of olfactory imagery therefore holds both theoretical and applied relevance, with implications for consumer neuroscience, food science, and the study of memory and multisensory cognition. A key methodological challenge is that primary olfactory structures reside in deep subcortical regions inaccessible to fNIRS, yet higher-order olfactory regions including the orbitofrontal cortex fall within its measurement depth. We therefore conducted a feasibility study evaluating whether functional near-infrared spectroscopy (fNIRS) can detect cortical signatures of olfactory mental imagery in these accessible regions. Thirty-five participants were recruited, of whom 30 were included in the final analysis. Participants generated visual or olfactory mental images in response to semantic cues (odor-related words) or perceptual cues (abstract color patterns) in a 2x2 factorial design. Cortical haemodynamic responses were measured using a 42-channel fNIRS array covering frontal, temporal, and parietal regions, with data analyzed using oxyhaemoglobin (HbO), deoxyhaemoglobin (HbR), and correlation-based signal improvement (CBSI), and group-level inference performed using surface-based statistical parametric mapping. Visual imagery produced robust bilateral frontoparietal activation across all chromophores, validating the experimental framework. Olfactory imagery produced modest bilateral orbitofrontal activation, detected exclusively in HbO measurements, providing tentative preliminary evidence that fNIRS can detect higher-order olfactory imagery-related cortical activity. Post-hoc chromophore analysis revealed striking discordance across conditions, with only visual imagery to words producing concordant HbO and HbR responses surviving CBSI correction, and deactivations detected exclusively in HbO. These findings offer qualified support for the feasibility of fNIRS in this domain whilst raising important questions about chromophore coupling dynamics during endogenously driven cognitive processes and the suitability of CBSI assumptions beyond conventional stimulus-driven paradigms.