Masaru Nakamura ^1* Takahiko Nagamine ²

• ¹ Kosekai-Kusatsu Hospital, Hiroshima, Japan
• ² Sunlight Brain Research Center, Yamaguchi, Japan

Depressive symptoms are particularly common and widespread in patients with bipolar disorder (BD) and major depressive disorder (MDD). Distinguishing between BD and MDD is an important clinical challenge, as their pharmacological treatment regimens are different. Functional near-infrared diffuse spectroscopy (fNIRS) has attracted the interest of scientists and clinicians in the fields of both brain function and clinical psychiatry. Its advantages of non-invasiveness, low cost, and high sensitivity make it even more attractive to clinicians, and fNIRS has been studied as a biomarker closely related to brain cortical function and symptoms of psychiatric disorders. In Japan, NIRS waveforms are used to differentiate depression symptoms. However, the agreement between psychiatric diagnoses and auxiliary diagnoses based on brain activity patterns obtained using the NIRS verbal fluency task (VFT) is not necessarily high. When using NIRS as an auxiliary diagnostic tool in clinical practice, it is necessary to consider what factors may interfere with diagnosis.We investigated the concordance rate between psychiatrists' diagnoses and auxiliary diagnoses based on NIRS in a verbal fluency task. Psychiatrists' diagnoses were made by trained psychiatrists using a semi-structured interview according to the MINI-International Neuropsychiatric Interview. In addition, we hypothesized that patient-related factors including disease duration, severity of psychiatric symptoms, psychotropic drugs and serum electrolytes may contribute to the discrepancy. Informed consent was provided in an opt-out format, and data were collected retrospectively from 118 BD patients and 123 MDD patients. This study was approved by the Ethics Committee of the Sunlight Brain Research Center, approval number SBRC-No.505. The results of the study showed that the concordance rates for BD and MDD diagnoses were 44.0% and 38.2%, respectively, with approximately half of the patients being classified as different disorders by NIRS. Patients with BD who could not be diagnosed by NIRS had statistically significantly lower serum sodium concentrations than those with BD who could be diagnosed by NIRS. Patients with MDD who could not be diagnosed by NIRS received higher antidepressant doses than those who could be diagnosed by NIRS. The multivariate logistic regression analysis revealed that higher antidepressant doses and lower sodium concentrations in patients with BD, and higher sodium concentrations in patients with MDD increased the discrepancy, respectively (Table 1). Therefore, we considered the possibility that serum electrolytes and antidepressant doses might independently affect neurovascular coupling. Abbreviations: BD = bipolar disorder, MDD = major depressive disorder, F = Female, M = Male, GAF = Global Assessment of Functioning, Cp eq = chlorpromazine equivalents, IMP eq = imipramine equivalents, DZP eq = benzodiazepine equivalents, Hb = hemoglobin, S-Na = serum sodium, S-K = serum potassium, FBS = fasting blood sugar