Stress impacts the way brain networks function, potentially reshaping their connectivity patterns and structural arrangement across different frequency ranges. This research utilizes the SAM 40 dataset to examine how stress alters brain network configurations in various electroencephalography (EEG) frequency bands, employing metrics of triadic interactions and network topology. EEG recordings were collected from 40 individuals engaged in cognitive exercises (Symmetric Mirror Images, Stroop Color-Word Tasks, Mental Arithmetic Tasks) and a relaxation state. Participants rated the stressfulness of tasks on a scale from 1 (minimal) to 10 (extreme), with tasks rated above 6 classified as high stress (HS) and those below 4 as low stress (LS). Functional connectivity was assessed using a method based on topographical profile similarity, termed high-order FC (tHOFC), across five frequency bands: Delta, Theta, Alpha, Beta, and Gamma. Studied metrics were triadic frequency (measured by triad counts |Ti|, i=0-3), total triad energy (Un), hub tendencies (TMHp for positive and TMHn for negative connections), and link counts (|P| for positive, |N| for negative links), with differences between stressed and relaxed states tested via permutation test. In the Alpha band, HS showed notable deviations from the relaxed state, with increased |T3|, Un, and |P|, alongside reduced |T1|, TMHn, and |N|. In the Beta band, HS displayed elevated |T3| and |P|, with decreased |N|. LS differed from the relaxed state only in the Alpha band, showing higher Un in frontal regions. Analysis of Un by triad type highlighted unique contributions, with HS and LS exhibiting increased Un(T1) and Un(T3) in frontal areas and diminished Un(T2) in temporal-parietal zones. Stress markedly reshapes brain network stability and structure, especially in the Alpha and Beta bands, with Un emerging as a key indicator for both HS and LS conditions.