Neurophysiological Effects of Daily Centrum Use on Mood and Well-Being: A Cognalyzer EEG Analysis in Adults Between 40-60

Relaxation Effect Level (REL) is a continuous, algorithmically derived index of relaxation computed from a 2-minute resting EEG segment recorded at 8 scalp electrodes (Fp1, Fp2, T3, T4, P3, P4, O1, O2; Fz ref; 256 Hz). After automated artifact rejection, 200 features are computed across canonical frequency bands, including absolute and relative power spectral density, cross-power spectral density, and connectivity metrics (phase coherence). A support vector machine trained on an independent dataset of EEG labelled with concurrent subjective relaxation ratings uses a fixed feature subset (mRMR/LightGBM selection) to output one score per segment, from -10 (least relaxed) to +10 (most relaxed); higher scores indicate greater relaxation. Model and pipeline were locked prior to enrollment. Each session yields 2 pre-task and 2 post-cognitive-load segments; the reported value is mean post-task REL minus mean pre-task REL within a session, and the two sessions will be compared.

COL is a continuous, algorithmically derived index of confusion computed from a 2-minute resting EEG segment recorded at 8 scalp electrodes (Fp1, Fp2, T3, T4, P3, P4, O1, O2; Fz ref; 256 Hz). After automated artifact rejection, 200 features are computed across canonical frequency bands, including absolute and relative power spectral density, cross-power spectral density, and connectivity metrics (phase coherence). A support vector machine trained on an independent dataset of EEG labelled with concurrent subjective confusion ratings uses a fixed feature subset (mRMR/LightGBM selection) to output one score per segment, from -10 (least confused) to +10 (most confused); higher scores indicate greater confusion. Model and pipeline were locked prior to enrollment. Each session yields 2 pre-task and 2 post-cognitive-load segments; the reported value is mean post-task COL minus mean pre-task COL within a session, and the two sessions will be compared.

AEL is a continuous, algorithmically derived index of arousal computed from a 2-minute resting EEG segment recorded at 8 scalp electrodes (Fp1, Fp2, T3, T4, P3, P4, O1, O2; Fz ref; 256 Hz). After automated artifact rejection, 200 features are computed across canonical frequency bands, including absolute and relative power spectral density, cross-power spectral density, and connectivity metrics (phase coherence). A support vector machine trained on an independent dataset of EEG labelled with concurrent subjective arousal ratings uses a fixed feature subset (mRMR/LightGBM selection) to output one score per segment, from -10 (least aroused) to +10 (most aroused); higher scores indicate greater arousal. Model and pipeline were locked prior to enrollment. Each session yields 2 pre-task and 2 post-cognitive-load segments; the reported value is mean post-task AEL minus mean pre-task AEL within a session, and the two sessions will be compared.

BRUMS self-report compared between baseline and post product-use visits, before and after cognitive tests.

VEL is a continuous, algorithmically derived index of mood valence computed from a 2-minute resting EEG segment recorded at 8 scalp electrodes (Fp1, Fp2, T3, T4, P3, P4, O1, O2; Fz ref; 256 Hz). After automated artifact rejection, 200 features are computed across canonical frequency bands, including absolute and relative power spectral density, cross-power spectral density, and connectivity metrics (phase coherence). A support vector machine trained on an independent dataset of EEG labelled with concurrent subjective mood ratings uses a fixed feature subset (mRMR/LightGBM selection) to output one score per segment, scores range from -10 to +10: negative values indicate negative (unpleasant) mood, positive values indicate positive (pleasant) mood. Model and pipeline were locked prior to enrollment. Each session yields 2 pre-task and 2 post-cognitive-load segments; the reported value is mean post-task VEL minus mean pre-task VEL within a session, and the two sessions will be compared.

AXL is a continuous, algorithmically derived index of anxiety computed from a 2-minute resting EEG segment recorded at 8 scalp electrodes (Fp1, Fp2, T3, T4, P3, P4, O1, O2; Fz ref; 256 Hz). After automated artifact rejection, 200 features are computed across canonical frequency bands, including absolute and relative power spectral density, cross-power spectral density, and connectivity metrics (phase coherence). A support vector machine trained on an independent dataset of EEG labelled with concurrent subjective confusion ratings uses a fixed feature subset (mRMR/LightGBM selection) to output one score per segment, from -10 (least anxious) to +10 (most anxious); higher scores indicate greater anxiety. Model and pipeline were locked prior to enrollment. Each session yields 2 pre-task and 2 post-cognitive-load segments; the reported value is mean post-task AXL minus mean pre-task AXL within a session, and the two sessions will be compared.