Decoding aroma perception of grilled lamb skewers: an EEG-MambaFusionNet framework integrating TDS and GC-IMS

This study investigates the dynamic evolution of aroma perception in grilled lamb skewers from raw to well-done stages and its corresponding neural signatures. Temporal Dominance of Sensations (TDS) and Gas Chromatography-Ion Mobility Spectrometry (GC-IMS) were integrated to capture sensory transitions from ‘mutton-distinctive’ to ‘charcoal-roasted’, ‘fatty’, ‘grilling-meat’ and finally ‘burnt’ aromas, along with associated volatile compound profiles. Concurrent electroencephalography (EEG) recordings showed that these aroma variations elicit distinct and progressively stronger neural responses. A critical time window was identified at 8–10 s post-inhalation, during which neural activity exhibited the highest discriminability across grilling stages. EEG-MambaFusionNet, a deep learning framework fusing neural, sensory, and chemical data, was developed to quantify this process, achieving a predictive accuracy of 0.925. Interpretation analysis indicated that prefrontal and parietal electrodes function as key hubs for hedonic and integrative aroma processing. These findings reveal the processing mechanism of brain for the aroma of grilled lamb skewers.