Introduction
Social cognition refers to cognitive processes, including social cue perception, inferring others’ thoughts or emotions, and managing emotional reactions or behavior to others (Arioli, Crespi, & Canessa, Reference Arioli, Crespi and Canessa2018). This ability is evolutionarily advantageous for individuals who live in groups, as it allows for enhanced protection against predators (Tomasello, Reference Tomasello2014). Social cognitive function is impaired in individuals with Bipolar disorder (BD), who may take longer to differentiate emotions, and exhibit lower accuracy compared to healthy controls (HCs), particularly when under time pressure (Gillissie et al., Reference Gillissie, Lui, Ceban, Miskowiak, Gok, Cao and McIntyre2022). For these individuals, their mood swings can further impair their ability to accurately interpret social cues and respond appropriately, leading to a vicious cycle of increasingly impaired social cognitive difficulties (Forgas, Bower, & Krantz, Reference Forgas, Bower and Krantz1984). We have previously demonstrated that individuals with BD exhibit social cognitive deficits, particularly in emotion recognition, across multiple studies (Chang et al., Reference Chang, Chang, Wu, Tsai, Lu, Chang and Tseng2024; Lee et al., Reference Lee, Huang, Chang, Kuo, Lu, Hsieh and Tseng2022; Tsai et al., Reference Tsai, Lu, Tseng, Chang, Wang, Yang and Chen2023a; Tsai et al., Reference Tsai, Chang, Wu, Huang, Chang, Lu and Tseng2023b). These deficits are closely linked to broader cognitive impairments (Tsai et al., Reference Tsai, Chang, Wu, Huang, Chang, Lu and Tseng2023b) and contribute to less favorable real-world outcomes, including interpersonal function (Lee et al., Reference Lee, Huang, Chang, Kuo, Lu, Hsieh and Tseng2022) and quality of life (Tamura et al., Reference Tamura, Carvalho, Leanna, Feng, Rosenblat, Mansur and McIntyre2022). Social cognition serves as a critical bridge between cognitive deficits and functional outcomes (Ospina et al., Reference Ospina, Nitzburg, Shanahan, Perez-Rodriguez, Larsen, Latifoglu and Burdick2018), and can be an important intervention target to enhance the quality of life for individuals with BD (Gillissie et al., Reference Gillissie, Lui, Ceban, Miskowiak, Gok, Cao and McIntyre2022).
Social cognition impairments in BD are underpinned by both neurofunctional and neurochemical changes in the brain. Several neurotransmitters, such as glutamate and GABA, and specific regions, such as the default mode network (DMN), salience network, and prefrontal cortex (PFC), have been implicated (Lopatina et al., Reference Lopatina, Komleva, Gorina, Olovyannikova, Trufanova, Hashimoto and Salmina2018). The PFC is a crucial regulator of social cognition, and disturbances in prefrontal microcircuitry contribute significantly to the pathophysiology of social deficits in psychiatric disorders. (Bicks, Koike, Akbarian, & Morishita, Reference Bicks, Koike, Akbarian and Morishita2015; Tso et al., Reference Tso, Burton, Lasagna, Rutherford, Yao, Peltier and Taylor2021). While mPFC is one of the critical hubs of the DMN, the manifestation of DMN in social cognitive deficits is also evident (Tso et al., Reference Tso, Burton, Lasagna, Rutherford, Yao, Peltier and Taylor2021).
A direct link between an impaired balance of excitatory and inhibitory neurotransmitters and changes in social behavior has been confirmed in recent studies in mice and humans (Lopatina et al., Reference Lopatina, Komleva, Gorina, Olovyannikova, Trufanova, Hashimoto and Salmina2018). The relationship between E/I balance in the PFC and social behavior was further supported by evidence showing that modulating the E/I balance in the medial prefrontal cortex (mPFC) of mice can rescue autism-like social behavior deficits (Selimbeyoglu et al., Reference Selimbeyoglu, Kim, Inoue, Lee, Hong, Kauvar and Deisseroth2017). In BD, previous research has identified excitatory (glutamate and glutamine) and inhibitory (GABA) neurometabolite alterations and their equilibrium by Proton magnetic resonance spectroscopy (1H-MRS) in mPFC/ACC (Scotti-Muzzi, Umla-Runge, & Soeiro-de-Souza, Reference Scotti-Muzzi, Umla-Runge and Soeiro-de-Souza2021; Sosa-Moscoso et al., Reference Sosa-Moscoso, Ullauri, Chiriboga, Silva, Haro and Leon-Rojas2022) and PCC/PCu, the hub of the DMN (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024).
Furthermore, studies have demonstrated a significant positive association between GABA and DMN deactivation; DMN deactivation was also correlated with working memory, suggesting functional relevance to its deactivation (Hu, Chen, Gu, & Yang, Reference Hu, Chen, Gu and Yang2013). Our research team has further identified an association between E/I imbalance within the DMN and executive function performance in individuals with euBD (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024). Furthermore, evidence suggests that neurocognitive impairments, particularly in executive function and memory, are closely linked to deficits in social cognition among individuals with severe mental disorders, including BD (Lancaster, Evans, Bond, & Lysaker, Reference Lancaster, Evans, Bond and Lysaker2003).
In light of the above findings, we hypothesize that the E/I ratio in the mPFC or PCC—key hubs of the DMN—is associated with impaired emotion recognition in individuals with euBD. Furthermore, we examine whether neurocognitive function mediates this association, providing insight into the potential pathway by which neurochemical imbalance impacts social cognitive performance.
Methods
Ethics statement
The study received approval from the Institutional Review Board for the Protection of Human Subjects at National Cheng Kung University Hospital. Patients were recruited from the hospital’s psychiatric outpatient department, while healthy volunteers were enlisted via online platforms and public advertisements. All participants provided written informed consent before enrollment.
Participants
Our study enrolled 49 euBD patients according to the Diagnostic Systematic Manual-5th edition. The BD patients were recruited from outpatient psychiatry clinics in the Department of Psychiatry, National Cheng Kung University Hospital, College of Medicine. Seventy-five HCs were enrolled from the community through advertisement. All participants, aged 18–65 years, underwent assessment by an attending psychiatrist using the Chinese adaptation of the Mini International Neuropsychiatric Interview (Sheehan et al., Reference Sheehan, Lecrubier, Sheehan, Amorim, Janavs, Weiller and Dunbar1998). The severity of depressive and manic symptoms was quantified with the 17-item Hamilton Depression Rating Scale (HDRS) and the 11-item Young Mania Rating Scale (YMRS). Euthymia was defined as achieving scores of 7 or lower on both the HDRS and YMRS scales (Zhang et al., Reference Zhang, Long, Ma, He, Luo, Bian and Xiang2018). The exclusion criteria were as follows: (i) a serious surgical condition or physical illness; (ii) major mental illnesses except BD for the BD patients; (iii) pregnancy or breastfeeding; (iv) substance abuse within the past 3 months, except tobacco use disorder; (v) previous use of any psychotropic agent in the HCs; and (vi) an organic mental disease, mental retardation, or dementia. This cohort partly overlaps with that described in our previously published work (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024).
Image parameters: proton magnetic resonance spectroscopy (1H-MRS)
Participants received imaging at the Mind Research and Imaging Center, National Cheng Kung University, utilizing a 3.0 Tesla MRI scanner (GE Discovery MR750, GE Medical Systems) with an 8-channel head coil for data acquisition. 1H-MRS was employed to evaluate excitatory (glutamate + glutamine, Glx) and inhibitory (GABA) neurometabolites, enabling the computation of E/I ratios. Glx levels were quantified via point-resolved spectroscopy (PRESS), while GABA concentrations were examined with MEGA-PRESS. The regions of interest, including the mPFC/ACC and PCC/PCu, were selected based on predefined anatomical coordinates (Hu et al., Reference Hu, Chen, Gu and Yang2013; Kegeles et al., Reference Kegeles, Mao, Stanford, Girgis, Ojeil, Xu and Shungu2012). Neurometabolites were quantified using LCModel, with a Cramér-Rao lower bound (CRLB) threshold of <20%. Partial volume correction accounted for tissue composition using segmentation and analysis tools (Gannet and MRSParVolCorr). This methodology has been previously described in our team’s prior publication (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024).
Social cognition and neuropsychological assessment
Non-verbal emotion recognition (DANVA-2-TW)
Emotional expressions, while partly universal, vary subtly across cultures, affecting the accuracy of nonverbal emotion recognition based on race, culture, and gender (Jack, Garrod, Yu, Caldara, & Schyns, Reference Jack, Garrod, Yu, Caldara and Schyns2012; Wickline, Bailey, & Nowicki, Reference Wickline, Bailey and Nowicki2009). To measure this accuracy in Han Chinese, we used the culturally adapted Diagnostic Analyses of Nonverbal Accuracy 2, Taiwanese Version (DANVA-2-TW). This version, based on the original DANVA-2 (Nowicki & Duke, Reference Nowicki and Duke1994), has been applied in Taiwanese clinical studies (Pan, Tseng, & Liu, Reference Pan, Tseng and Liu2013). The DANVA-2-TW includes 60 facial photos and 60 voice recordings depicting happiness, sadness, anger, fear, and neutral emotions, with 12 examples per emotion. Photos were shown on a 1024 × 768 screen, and audio was played via headphones. Accuracy was calculated as the proportion of correct responses per (Pan, Tseng, & Liu, Reference Pan, Tseng and Liu2013) emotion, ranging from 0 (completely inaccurate) to 1 (completely accurate).
Wisconsin Card-Sorting Test (WCST)
An experienced clinical neuropsychologist administered the WCST, which consisted of 64 cards. All index definitions followed the guidelines outlined in the WCST manual (Heaton, Chelune, Talley, Kay, & Curtiss, Reference Heaton, Chelune, Talley, Kay and Curtiss1993). Participants completed a computerized version of the WCST, where they matched response cards to four stimulus cards based on color, shape, or quantity, receiving feedback after each attempt. The patients were not informed about the dimensions beforehand. Once patients successfully sorted 10 cards within a given category, they were instructed to reclassify them based on a new category. Performance on the WCST was evaluated using the number of completed categories and perseverative errors.
Statistical analysis
Chi-squared test was used to compare the demographic and clinical characteristics of the participant groups for dichotomous variables. Differences between the BD and HC groups were assessed through independent t-tests. Likewise, MRS variables were examined using the same method, with additional adjustments for age and sex performed via analysis of covariance (ANCOVA) to account for potential confounding factors.
To explore the relationship between emotion recognition and the E/I ratio in BD, Pearson correlation analysis was performed to assess the association between MRS variables and DANVA-2-TW within the BD group. A significance threshold of p < 0.05 was applied to all statistical tests. Mediation effects were evaluated using the Sobel Test Calculator. All statistical analyses were conducted using IBM SPSS Statistics, version 17 (IBM Corporation, Armonk, NY, USA).
Results
Demographic data and group differences
We compared the BD and HC groups in terms of age, sex, educational years, clinical mood symptoms, emotion recognition ability, neurocognition, and neurometabolite variables (Table 1). The BD group had a higher proportion of females (p = 0.001) and was on average older (38.31 ± 13.22, p < 0.001) than the HC group. Hence, age and sex were controlled when comparing emotion recognition ability, neurocognition, and neurometabolite variables between the BD and HC groups. The BD group had higher subsyndromal depressive and manic symptoms compared to the HC group, as measured by the HDRS (p = 0.014) and YMRS (p < 0.001) scales, respectively.
Demographic and clinical data
As shown in Table 2, the BD group had lower abilities in recognizing happy (0.73 ± 0.16, p = 0.002), sad (0.57 ± 0.19, p = 0.021), angry (0.55 ± 0.19, p = 0.036) and overall emotion (0.60 ± 0.19, p = 0.020) compared to the HC group. Moreover, the BD patients performed worse in the WCST categories completed test than the HCs (2.73 ± 1.69, p = 0.002) but not in perseverative error (11.20 ± 10.85, p = 0.433).
Neuroimaging and neuropsychological variables
Abbreviations: MRS, magnetic resonance spectroscopy; Glx: glutamate + glutamine signal complex; PCC: posterior cingulate cortex; mPFC/ACC: medial prefrontal cortex/anterior cingulate cortex; GABA: gamma-aminobutyric acid; CRLB: Cramér-Rao lower bound; DANVA-2-TW: diagnostic analyses of nonverbal accuracy, Taiwanese version.
a Controlling for age and sex.
With regard to MRS variables, the BD patients showed higher glutamate complex levels in both of the PCC (9.60 ± 0.94, p = 0.001) and mPFC/ACC (9.91 ± 2.94, p = 0.011). The E/I ratio was higher in mPFC/ACC and remained a trend of significance after controlling for age and sex (p = 0.099).
The association of the E/I ratio and emotion recognition
A negative association between emotion recognition (DANVA-2-TW total score) and the glutamatergic-GABAergic balance was observed in the mPFC/ACC of the BD patients (r = −0.30, p = 0.034) (Figure 1). However, this relationship was not seen in the HC group (r = −0.04, p = 0.732).
Summary of study design, correlation, and mediation analysis.
Note: euBD patients (n = 49) and HCs (n = 75) underwent 3 T 1H-MRS and emotion recognition assessment (DANVA-2-TW). In BD, a negative correlation was found between prefrontal E/I ratio and emotion recognition. Mediation analysis showed WCST performance partially mediated this association (Sobel Z = −2.657, p = 0.007). Abbreviations: euBD, euthymic bipolar disorder; HCs, healthy controls; 1H-MRS, Proton magnetic resonance spectroscopy; DANVA-2-TW, Diagnostic Analysis of Nonverbal Accuracy 2-Taiwan version; E/I, excitatory/inhibitory; WCST, Wisconsin Card Sorting Test; mPFC/ACC, medial prefrontal cortex/anterior cingulate cortex.
After including WCST categories completed as a control variable in the correlation analysis, the correlation between the DANVA total score and the mPFC/ACC E/I ratio of the BD patients was no longer significant (p = 0.873). Hence, we proposed a model of the mediating effect of neurocognitive ability in emotion recognition. We tested the proposed model with the Sobel analysis (Figure 1) and found a significant mediating effect of neurocognition (WCST) on E/I balance in the mPFC/ACC and emotion recognition (Z = −2.657, p = 0. 007).
Discussion
Building on our previous research demonstrating an association between E/I imbalance within the DMN—particularly in the PCC and mPFC/ACC—and executive function in euBD (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024), the current study extends this line of investigation into the social cognitive domain. Specifically, we examined whether the same E/I imbalance is also related to emotion recognition and whether this association is mediated by executive function. To the best of our knowledge, this mediation model has not been previously examined in bipolar disorder. Our findings offer preliminary evidence suggesting a potential neurocognitive pathway linking neurometabolic alterations with social cognitive functioning. To the extent of our knowledge, this research represents the first investigation into the correlation between E/I imbalance and emotion recognition in BD. We found that the E/I ratio in the mPFC/ACC negatively predicted emotion recognition ability in BD, the negative association was observed when the WCST categories completed were added to the model. The Sobel test proved that a mediating effect of cognitive flexibility on the correlation between the E/I ratio and emotion recognition in the mPFC/ACC. Consistent with the hypothesis, the results support the association between higher E/I balance within the DMN and poorer social cognition. Additionally, neurocognitive functioning, particularly executive function, was found to mediate this relationship.
Metabolite alterations were also mentioned in cross-sectional MRS studies comparing BD patients with HCs (Sosa-Moscoso et al., Reference Sosa-Moscoso, Ullauri, Chiriboga, Silva, Haro and Leon-Rojas2022), and the regions most commonly affected in BD include the prefrontal cortex, anterior cingulate cortex, and basal ganglion (Sosa-Moscoso et al., Reference Sosa-Moscoso, Ullauri, Chiriboga, Silva, Haro and Leon-Rojas2022). The neocortex is primarily composed of glutamatergic excitatory pyramidal neurons, making up the majority of its neuronal population (Somogyi, Tamás, Lujan, & Buhl, Reference Somogyi, Tamás, Lujan and Buhl1998). Approximately 20% of cortical neurons are GABAergic inhibitory interneurons (Tatti, Haley, Swanson, Tselha, & Maffei, Reference Tatti, Haley, Swanson, Tselha and Maffei2017). The brain’s information processing relies on a delicate equilibrium between the excitatory signals generated by glutamatergic pyramidal neurons and the inhibitory signals from GABAergic interneurons. The interaction between the two contrasting processes of excitation and suppression can impact various neural circuits and brain function (Liu et al., Reference Liu, Ouyang, Zheng, Mi, Zhao, Ning and Guo2021; Sohal & Rubenstein, Reference Sohal and Rubenstein2019) and has been suggested as a framework for understanding the pathophysiology of neurodevelopmental and neuropsychiatric disorders (Liu et al., Reference Liu, Ouyang, Zheng, Mi, Zhao, Ning and Guo2021).
The regional E/I balance has been shown to predict DMN function, which is associated with goal-directed task performance (Gu, Hu, Chen, He, & Yang, Reference Gu, Hu, Chen, He and Yang2019). Effective suppression of DMN activity is essential for goal-directed cognition, potentially by reducing goal-irrelevant processes such as mind wandering. This framework aligns with our series of studies, in which we found that the E/I ratio in the mPFC/ACC negatively predicted both emotion recognition and executive function in individuals with BD (Tseng et al., Reference Tseng, Wu, Chang, Lu, Chang, Hsu and Chen2024). These findings further support the notion that an optimal E/I balance in the mPFC/ACC is critical for regulating DMN function, thereby influencing cognitive and socio-emotional processes. Studies have found that individuals with schizophrenia and their siblings show impairments in both facial emotion recognition and executive functions, with correlations between these deficits (Jaracz, Grzechowiak, Raczkowiak, & Rybakowski, Reference Jaracz, Grzechowiak, Raczkowiak and Rybakowski2011; Yang et al., Reference Yang, Zhang, Li, Heeramun-Aubeeluck, Liu, Huang and Lu2015). Similar findings were observed in BD patients (David, Soeiro-de-Souza, Moreno, & Bio, Reference David, Soeiro-de-Souza, Moreno and Bio2014). Higher emotional intelligence has been associated with better WCST performance, indicating a link between emotional competencies and prefrontal cortex functions (Alipour, Arefnasab, & Babamahmoodi, Reference Alipour, Arefnasab and Babamahmoodi2011). These findings underscore the interconnectedness of executive functions and emotional processing across different populations, aligning with the results of our own research. The mediating effect of neurocognition on emotional recognition was demonstrated through a model, where bottom-up perceptual processes are connected to social cognitive skills, such as facial recognition. These abilities are consistently influenced by neurocognitive functions, collectively contributing to the ability to recognize emotions (Ventura, Wood, Jimenez, & Hellemann, Reference Ventura, Wood, Jimenez and Hellemann2013). Extending our findings, we propose an explanation linking neurometabolic imbalance to impaired emotion recognition through E/I dysregulation in the PFC. Specifically, we suggest that an increased E/I ratio within the mPFC/ACC may disrupt the balance between excitatory and inhibitory signaling in cortical microcircuits—an equilibrium essential for neural decorrelation and efficient information processing (Chini, Pfeffer, & Hanganu-Opatz, Reference Chini, Pfeffer and Hanganu-Opatz2022). This imbalance has been shown in animal studies to impair both cellular computation and social behavior (Yizhar et al., Reference Yizhar, Fenno, Prigge, Schneider, Davidson, O’Shea and Deisseroth2011), indicating functional consequences that span molecular, circuit, and behavioral levels. At the cognitive level, an elevated E/I ratio may interfere with evidence accumulation and decision-making, leading to impulsive judgments that overweight early cues (Lam et al., Reference Lam, Borduqui, Hallak, Roque, Anticevic, Krystal and Murray2017). Computational modeling supports this notion, demonstrating that even subtle elevations in E/I ratio can destabilize decision dynamics and promote impulsive selections (Murray & Wang, Reference Murray, Wang, Anticevic and Murray2018). Such effects are especially relevant to emotion recognition tasks, which require the dynamic integration of subtle facial or vocal cues over time. As Freeman et al. emphasize, person perception involves a continuous interplay between bottom-up and top-down information (Freeman, Johnson, Adams, & Ambady, Reference Freeman, Johnson, Adams and Ambady2012). Disruption of this temporal integration—potentially due to early sensory bias driven by E/I imbalance—may thus impair the accurate decoding of emotional signals (Freeman et al., Reference Freeman, Johnson, Adams and Ambady2012; Lam et al., Reference Lam, Borduqui, Hallak, Roque, Anticevic, Krystal and Murray2017). Together, these findings provide a plausible mechanistic framework linking prefrontal neurometabolic imbalance to socio-cognitive dysfunction, mediated through impaired executive functioning.
There were some limitations of our study. First, the number of subjects included in the study was small, so caution is needed to generalize these findings to BD patients as a whole. Larger samples are needed to obtain more reliable results. Moreover, a potential limitation of this study is the lack of control for years of education, specifically in the within-group analyses. While education level is often considered a proxy for premorbid intellectual functioning and is known to influence performance on tasks of emotion recognition and executive function, its interpretation within clinical samples requires caution. In individuals with bipolar disorder, educational attainment may not solely reflect premorbid cognitive capacity, but also the cumulative impact of illness-related factors such as early onset, chronicity, or cognitive decline. As such, adjusting for education in analyses could lead to overadjustment and inadvertently remove variance that is meaningful for understanding illness-related neurobiological changes.
In the present study, our primary analyses were conducted without controlling for education in order to preserve the interpretability of illness-related variance. Nevertheless, to address potential confounding, we performed sensitivity analyses controlling for education, and these results—presented in the Supplementary Materials—showed attenuation of the correlation between E/I balance and emotion recognition.
Higher educational attainment has been implicated in cognitive resilience and structural brain modifications, including increased gray matter volume and metabolic activity in the anterior cingulate cortex, lingual gyri, and precuneus (Arenaza-Urquijo et al., Reference Arenaza-Urquijo, Landeau, La Joie, Mevel, Mézenge, Perrotin and Chételat2013; Eisenberg et al., Reference Eisenberg, London, Matochik, Derbyshire, Cohen, Steinfeld and Galynker2005). However, the relationship with neurometabolite concentrations remains inadequately characterized. While limited evidence suggests that greater educational attainment is associated with elevated whole-brain N-acetylaspartate levels—an established marker of neuronal integrity—particularly in younger adults (Glodzik et al., Reference Glodzik, Wu, Babb, Achtnichts, Amann, Sollberger and Gonen2012), the precise nature and extent of this association remain unclear. Consequently, the years of education which was not controlled. Additionally, there are some subtype features and multiple dimensions within BD (Angst, Reference Angst2007), and these may contribute to different neurometabolite characteristics. On the other hand, MRS measurement variability depends on many technical factors, including hardware, acquisition parameters, data quality, and data analysis (Harris et al., Reference Harris, Amiri, Bento, Cohen, Ching, Cudalbu and Bartnik-Olson2022). Additional biological challenges with MRS are that metabolite levels are linked to the tissue fraction in the voxel; moreover, neurometabolites may change with age (Harris et al., Reference Harris, Amiri, Bento, Cohen, Ching, Cudalbu and Bartnik-Olson2022). Concerning those MRS limitations, studies evaluating the roles of neurometabolites need careful interpretation and take MRS harmonization into consideration. Finally, several animal models have proven that mood stabilizers exert therapeutic effects by altering the synaptic E/I balance (Lee, Zhang, Kim, & Han, Reference Lee, Zhang, Kim and Han2018), which may have confounded our results. Further comparison of subgroups, such as medicated and non-medicated BD, BDI/II patients, or patients with psychotic features or not, in a larger sample size is warranted in the future. Lastly, this cross-sectional study had limitations in identifying causal inference and establishing neurobiological pathways.
In summary, this study provides preliminary evidence for a possible mechanistic pathway in which E/I imbalance may influence social cognition indirectly through executive functioning, specifically affecting emotion recognition in euBD. This finding suggests that prefrontal executive function, particularly cognitive flexibility, could play a mediating role linking neurobiological dysregulation to social cognitive performance, addressing an area that has received limited attention in the literature.
These results may also have practical implications. They indicate that social cognitive interventions might be enhanced by including components that support executive functions, such as cognitive flexibility and working memory, in addition to conventional social scenario training. Such an integrated approach could potentially increase the effectiveness of psychosocial interventions and contribute to improved functional outcomes.
Abbreviations
- 1H-MRS
-
Proton magnetic resonance spectroscopy
- ANCOVA
-
analysis of covariance
- CRLB
-
Cramér-Rao lower bound
- DANVA-2-TW
-
Diagnostic Analysis of Nonverbal Accuracy 2-Taiwan version
- DMN
-
default mode network
- E/I
-
excitatory/inhibitory
- euBD
-
Euthymic bipolar disorder
- Glx
-
glutamate/glutamine
- HCs
-
healthy controls
- HDRS
-
Hamilton Depression Rating Scale
- mPFC/ACC
-
medial prefrontal cortex/anterior cingulate cortex
- PCC
-
posterior cingulate cortex
- PRESS
-
point-resolved spectroscopy
- WCST
-
Wisconsin Card Sorting Test
- YMRS
-
Young Mania Rating Scale
Acknowledgements
We sincerely appreciate all participants for their involvement in this study. Our gratitude also extends to the Mind Research and Imaging Center (MRIC) at National Cheng Kung University for providing consultation and access to research instruments. The MRIC is funded by the National Science and Technology Council.
Author contribution
Huai-Hsuan Tseng, the corresponding author, led the study’s conception and drafted the research protocol. Hui Hua Chang, Yen Kuang Yang, Po See Chen, and Yu-Lien Huang contributed to refining the study design. Statistical analyses were conducted by Chih-Yu Chang, Shyh-Yuh Wei, and Ren-Yi Lin. The initial manuscript draft was prepared by Cheng Ying Wu. Data collection was coordinated by Ying Tsung Tsai, Tsung-Hua Lu, Po See Chen, and Huai-Hsuan Tseng. All authors were actively involved in data interpretation and provided critical revisions to enhance the manuscript.
Funding statement
This study was financially supported by the Ministry of Science and Technology, Taiwan (NSTC 113-2314-B-006-071-MY3, MOST 107–2314-B-006-082, MOST 107–2320-B-006-016, MOST 107–2628-B-006-005, MOST 108–2628-B-006-004, MOST 109–2628-B-006-004, MOST 110–2314-B-006-061-MY3, and MOST 110–2320-B-006-022). The funding bodies had no influence over any aspect of the research, including study design, data collection, analysis, result interpretation, manuscript preparation, or the decision to submit for publication.
Competing interests
The authors declare no financial conflicts of interest.
Open access
