Nina Favor

Portable breath-sensing platforms offer a practical approach for collecting physiological and environmental information outside controlled laboratory settings, but their interpretation is complicated by inter-individual variability and sensor-related effects. This pilot study evaluated the BreathLogix VOC Series prototype, a portable multi-sensor platform incorporating gas-sensitive, carbon-dioxide, temperature, and environmental channels, during routine collegiate athletic activities. A total of 421 breath recordings were obtained from 209 participant ID codes, including 175 pre-exercise and 246 post-exercise recordings. Continuous sensor signals were converted into aggregated sample-level features and analyzed using descriptive statistics, interquartile-range comparisons, Pearson correlation analysis, and K-means clustering. Group-level median responses remained relatively stable across exercise conditions, whereas response variability changed more noticeably. Humidity showed the largest post-exercise increase in dispersion, followed by the hydrogen- and methane-sensitive channels, while the acetone-sensitive response showed reduced variability. Correlation analysis identified strong relationships among selected channels, consistent with possible cross-sensitivity and environmental influence. Clustering revealed two overlapping sensor-derived breath-signature groups primarily separated by acetone-sensitive response; a sex difference in cluster composition was not retained when each participant was counted once. These findings demonstrate the feasibility of field-based multi-channel breath acquisition and highlight the value of variability-centered analysis. Further analytical validation, calibration assessment, and longitudinal reproducibility testing are required.