Bipolar disorder (BD) is characterized by the recurrence of depression and mania, with depressive symptoms being the initial presentation in approximately 60 % of cases (Kapczinski et al., 2016; Kupka and Hillegers, 2022). A pivotal criterion in diagnosing BD is the identification of (hypo)manic episodes, a requirement for a definitive diagnosis. However, the challenge lies in the early stages of the disorder, where only 20 % of individuals presenting with depressive episodes are accurately diagnosed within the first year of seeking treatment (Goldberg et al., 2001), and the average delay in diagnosis was 6.46 years (Lublóy et al., 2020). This significant delay not only biases treatment approaches but also inflicts considerable mental and economic distress on patients and their families (Duffy et al., 2010), underscoring the urgency for early identification and intervention. The progressive nature of BD suggests a prelude of subthreshold manic symptoms before the emergence of full-blown manic episodes (Kupka and Hillegers, 2022). Between 30 %–55 % of individuals with depressive episodes may exhibit subthreshold manic symptoms, with over a third of adolescent depressions potentially evolving into BD (Angst et al., 2018; Judd et al., 2012), hinting a critical opportunity for early identification and intervention, especially among adolescents with major depressive disorder (MDD) exhibiting subthreshold manic symptoms, who are considered at high risk for BD (Faedda et al., 2019; Kupka and Hillegers, 2022; Lublóy et al., 2020; Salazar de Pablo et al., 2023; Salvatore et al., 2014; Waszczuk et al., 2022). The current characterization of subthreshold manic symptoms exhibits heterogeneity, often encompassing a wide array of criteria (Liu et al., 2022; Yang et al., 2016; Correll et al., 2014). Such clinical variability in defining subthreshold manic symptoms contributes to inconsistencies in neurobiological research findings, thereby impeding the processes of early detection and intervention. Consequently, there emerges a compelling necessity to delineate a more precise and efficacious index for subthreshold manic symptoms (Merikangas et al., 2014). Recent longitudinal clinical studies advocate for a definition of subthreshold mania as the presence of at least two manic symptoms, accumulating for a minimum of 4 h per day for 4 days in total. This refined definition facilitates more effective and feasible neuroimaging and cognitive function investigations during the prodromal stage of BD (Beekman et al., 2022; Fiedorowicz et al., 2011; Geller et al., 2004; Papachristou et al., 2017).

Neuroimaging research has begun to elucidate neurobiological alterations associated with subthreshold BD, revealing changes across structural connectivity, region of interest (ROI)-based resting-state functional MRI (rs-fMRI), and task-related functional neuroimaging. Notably, a decreasing trend in interregional connectivity involving nodes within the right prefrontal and temporal lobes has been observed among individuals at high genetic risk for BD displaying mild symptoms (Bora et al., 2021). ROI analysis demonstrated increased DMN connectivity in MDD with mixed features (Liu et al., 2022), whereas task-based neuroimaging has shown heightened activation in frontal regions during emotional inhibition conditions in MDD patients with manic symptoms (Cha et al., 2021). Collectively, these studies delineate distinct neuroimaging abnormalities in individuals at high risk for BD, including whose with subthreshold manic symptoms or syndromes. However, the diversity in methodologies remains a significant obstacle in pinpointing consistent biological markers in this group (Zhang et al., 2016).

With the advancement of brain imaging techniques, graph theoretical analysis has become pivotal in modeling the human brain's functionally intricate network, offering insights into the roles of central regions through nodes and edges (Goldman et al., 2022). The global brain connectivity (GBC) metric, leveraging a data-driven rs-fMRI strategy, is instrumental in identifying hub-like regions. It provides voxel-based analyses, elucidating the central roles within the functional connectome and effectively overcoming the challenges posed by seed-based and parcellation-based connectome analyses (Abdallah et al., 2017; de Reus and van den Heuvel, 2013). However, despite these advancements, the application of GBC methodology in neurobiological research, particularly among individuals in the high-risk population for BD, remains underexplored based on this criterion.

As is a key aspect of cognitive performance, processing speed that is impaired in BD is often considered a promising cognitive endophenotype (Dobri et al., 2023). Patients at high risk for BD exhibiting subthreshold manic symptoms or syndromes demonstrate neurocognitive imaging mechanisms that are independent of genetic factors. However, research showing processing speed impairment in the offspring of individuals with BD who exhibit subthreshold syndromes is very limited (Lin et al., 2018; Nguyen et al., 2017; Zhang et al., 2019). The GBC alterations in adolescent patients with MDD experiencing subthreshold mania, as well as their relationship to the impairment in processing speed before the onset of full-blown BD, remain unclear.

In this study, adolescents with MDD experiencing subthreshold mania (SubMD) are considered at high risk for BD, while those without subthreshold manic symptoms (nSubMD) are classified as cases of pure depression (Fornaro et al., 2013). This study aims to elucidate GBC differences between SubMD and nSubMD patients and to explore the relationship between GBC alterations and processing speed, shedding light on the neurocognitive profiles of the clinical high-risk population for BD. We tentatively propose that global brain functional connectivity in SubMD adolescents differs from that in the nSubMD group, especially within prefrontal regions implicated in emotions and cognition. Moreover, we suggest that SubMD patients might exhibit processing speed impairments, potentially linked to prefrontal functional connectivity.