According to the current Diagnostic and Statistical Manual of Mental Disorders (DSM-5), sub-threshold bulimia nervosa (sub-BN) connotes lower duration or symptom frequency than full-threshold BN, which may also result in a similar degree of distress and functional impairment. Specifically, sub-BN has at least two uncontrollable binge-eating episodes and compensatory behavior episodes (i.e., self-induced vomiting, laxatives or diuretic use, fasting, and excessive exercise to compensate for overeating) per month for at least three months, as well as weight and shape concerns (American Psychiatric Association, 2013; Stice & Desjardins, 2018). In line with these core features, individuals with bulimia symptoms present with negative self-referential thoughts, which contribute to the maladaptive pattern of disordered eating behaviors, such as food restriction, binge eating, and purging. There is evidence that individuals with sub-BN had a higher prevalence of mood disorders (major depression and dysthymia) and also reported significantly more anxiety symptoms compared with individuals reporting no eating disorders (Touchette et al., 2011). However, previous studies have mostly investigated the neurobiological underpinnings of BN in clinical conditions, the structural and functional neural bases involved in sub-BN are relatively sparse or even unclear. As such, the identification of brain-based abnormalities which underpins and maintains sub-BN symptomatology is conducive to the prevention of developing BN and advance novel treatment initiatives.

Generally, the urge to binge might be mediated by hyperactivity of brain regions in reward circuit and hypoactivation of executive control networks (Seitz et al., 2016; Van den Eynde et al., 2010), therefore resulting in altered reward sensitivity and food-related attentional bias (Donnelly et al., 2018; Kessler et al., 2016). However, many imaging studies examined neural processes in those with BN and observed inconsistent findings(Cyr et al., 2017; Westwater et al., 2018; Wallace et al., 2019).

Voxel-based morphometry studies assessing structural characteristics of BN/non-clinical BN are relatively sparse, and are characterized by mixed findings in gray matter volume (GMV). For instance, studies have observed greater GMV of medial/lateral orbitofrontal cortex (OFC) and ventral/dorsal striatum in BN patients compared to healthy controls (HCs) (Schafer et al., 2010; Frank et al., 2016), which were part of the neural reward system and have linked with appetite regulation, and food reward processing (Burger & Stice, 2011). While Westwater et al. (2018) observed that increased bulimia symptoms correlated with reduced structural connectivity of left OFC and middle temporal cortex (MTG). Furthermore, it has been shown that a significant decrease in the inferior frontal gray matter of BN patients, which plays a role in control process (Hoffman et al., 1990). Yet in non-clinical samples, one recent study exhibited that no regional differences in cortical thickness were observed in adults who engage in binge episodes, compared to adults not engaging in binge episodes (Abdo et al., 2020). In contrast, another non-clinical study revealed greater GMV in the dorsolateral prefrontal cortex (dlPFC) was found in adults who report binge eating, a region which has previously been associated with disinhibited eating (Yao et al., 2016).

By a functional perspective, Lee et al. (2013) indicated that BN showed stronger synchronous activity between the dACC and medial OFC. Widespread alterations in these regions raise important question around the context specificity of reward circuit abnormalities. In addition, cumulative evidence demonstrated that fronto-striatal circuits are aberrant in BN (Celone et al., 2011; Cyr et al., 2016; Skunde, 2016). However, the findings are similarly mixed, both enhanced and reduced activation patterns have been detected. For instance, there is evidence revealed that diminished activity in nodes of the inhibitory control network, such as the dlPFC, ventral medial PFC and inferior frontal gyrus (IFG) (Bohon and Stice, 2011; Marsh et al., 2009; Neveu et al., 2018), suggesting that impaired recruitment of inhibitory control network may underpin an impaired ability to inhibit food intake in BN. Paradoxically, Stopyra et al. (2019) reported that individuals with BN demonstrated hyperactivation in the dlPFC within the executive network, as compared to those with binge eating disorder (BED). These findings suggest disrupted control system functioning in individuals with BN. However, the neural basis specifically underlying non-clinical populations remains unclear.

In addition to the observations above, the neural circuits that relate to self-referential thoughts of body shape/weight were also found in bulimia patients. Body image disturbance is central to BN, and flawed cognitions about shape/weight, coupled with societal pressures for thinness, lead individuals with BN to preferentially attend to negative aspects of their bodies, which could perpetuate binging and purging behaviors. In response to eating disorder-related words presentation, BN patients showed hypoactivation in occipital regions including parietal areas and precuneus (Pringle et al., 2011), which are implicated majorly in self-referencing processing (Kircher et al., 2000). Also when comparing their bodies with those of others, women with BN compared with HCs show decreased activation of precuneus (Mohr et al., 2011; Van den Eynde et al., 2013). While Domakonda et al. (2018) implied that individuals with BN exhibited hyperactive connectivity between self-referential related network and attention system and was associated with BN severity, reflecting abnormal engagement of attentional processes, perhaps related to their excessive attention to self-referential thoughts about body shape/weight. Moreover, from the perspective of cognitive behavioral models of BN, BN is characterized by low self-esteem that is linked with individuals’ high levels of body weight and shape concerns (e.g., negative physical self), thereby contributing to the development and maintenance of BN. It, thus, is interesting to explore how the neural basis underlying sub-BN relate to individual behavior characteristics.

The current study investigated between-group difference of the neuroanatomical abnormalities (GMV) and the spontaneously (or intrinsically) brain activity in a non-clinical sample, as to further contribute to prevent or decrease the prevalence rates of clinical BN. Specifically, VBM analysis and degree centrality (DC) analysis were used to examined aberrant changes in GMV and spontaneous brain activity in individuals with sub-BN, respectively. Then, regarding the resultant brain regions above as seeds, we further examined group difference in static functional connectivity (FC). In addition, several exploratory correlational analyses were applied to discover the proper brain-behavior associations.