Hampered gamma oscillations induced by sad emotion underlying suicide attempt in major depressive disorder

Aim Major depressive disorder (MDD) is associated with high suicidality, especially for those with suicide attempt (SA). Although impaired oscillatory activity has been previously reported in patients with SA, little is known about precise temporal–spatial variability of its neural dynamics. To solve this, the current study probed the spectral power and network interactions underlying SA in MDD. Methods The present study recruited 104 subjects including 56 subjects with MDD (30 with SA and 26 without SA) and 48 healthy controls, who performed sad expressions recognition task during magnetoencephalography (MEG) recording. By investigating source‐reconstructed MEG‐data, brain states representing different task stages were estimated from a Hidden Markov model. Spectrum power and network connectivity were compared via Gaussian Mixture Models, and fractional occupancy (FO) of states were compared via an independent F ‐test. Results Brain states were corresponding to various frequencies (theta/beta/low gamma/ high gamma). In low gamma band (35–45 Hz), the early visual state exhibited increased activation and hyper inter‐network connectivity between visual regions and the limbic system, while the middle fronto‐parietal state exhibited attenuated activation and decreased intra‐network connectivity within fronto‐parietal regions in SA group. Crucially, FO values of these two states were significantly correlated with the suicide risks. Conclusions Suicide behavior of patients with MDD was significantly associated with aberrant oscillations in low gamma band. Elevated oscillations in occipital cortices and attenuated oscillations in fronto‐parietal cortices were significantly associated with SA. Manifesting sadness indulging and reckless decision‐making, the hampered temporal characteristics could help explain the neural‐electric basis of SA.