Cerebellar neural markers of susceptibility to social isolation and positive affective processing

Humans are social species and the social ties that humans form with each other have an evolutionary value to enhance survival (Cacioppo et al. 2014). If individuals lack the social ties and are socially isolated, they often show problems with physical and mental health, such as increased cardiovascular health risk (Caspi et al. 2006), immune dysregulation (Pressman et al. 2005) and psychological distress (Levula et al. 2017). Conversely, an extensive social network can be a protective factor for dementia and cognitive function decline (Bennett et al. 2006). It was proposed that loneliness, the subjective negative feeling of the perception of being socially isolated, is key in explaining the detrimental impact of social isolation, such as the onset and severity of major depression (Cacioppo et al. 2010, 2015).

Although smaller social network size of an individual is often associated with higher loneliness feeling, there exists substantial individual difference in perception of loneliness even with the same level of objective social connection. Thus, there could be discordance between subjective (i.e., loneliness) and objective social connectedness, since it is not uncommon that people can still feel lonely despite having a rich social life with extensive interpersonal connections, or may not feel lonely even when living a solitary life. Limited studies have purposefully studied the individual difference in the discrepancy between social isolation and loneliness (McHugh et al. 2017), and its association with people’s cognitive and neural functions. In particular, it is postulated that susceptibility to social isolation is closely related to altered attentional neural mechanisms, with loneliness being linked with negative affective attentional biases. Lonely individuals showed reduced tendency to use positive thinking to alleviate their negative emotions (Hawkley et al. 2009). Young lonely adults also reported higher levels of trait neuroticism, which were in turn associated with increased left dorsolateral prefrontal cortex gray matter volume, suggesting altered emotion regulation functions (Kong et al. 2015).

While affect regulation has been primarily studied in relation to the functions of the prefrontal-limbic circuitries, with the lateral and medial prefrontal cortex exerting top–down regulatory influence over the subcortical limbic-striatal areas (Wu et al. 2018), the role of other brain regions that have also been implicated in both cognitive processing and emotion regulation, such as the cerebellum, is less known. The classic view of cerebellar function is that this region is primarily involved in sensori-motor coordination functions (Grodd et al. 2001). However, more recent literature suggests that the cerebellum is also critical to social and affective processing and regulation (Van Overwalle et al. 2014; Hoche et al. 2016; Guell et al. 2018), and damage to the cerebellum would lead to the ‘cerebellar cognitive affective syndrome’, characterized by deficits in executive function, disinhibition and blunted affect (Schmahmann and Sherman 1998). The cerebellum consists of ten lobules, with lobules I–V forming the anterior cerebellar lobe, lobules VI–IX forming the posterior cerebellar lobe, lobule X forming the flocculonodular lobe and the vermis dividing cerebellum medially (Molliver and O’Hearn 2001). Specifically, the vermis in the medial part of cerebellum may be particularly implicated in affective processes, and damage to this region was commonly observed in patients with affective disturbances, while the (lateral) posterior cerebellum may be mainly involved in higher cognitive functions (Exner et al. 2004; Stoodley and Schmahmann 2010; Schmahmann 2014). Patients with major depressive disorder showed reduced gray matter volume in the vermis (Beyer and Krishnan 2002), and greater functional activations within the cerebellum, along with other regions, were found when participants were viewing social rather than non-social scenes (Powers et al. 2013). According to recent evidence, the cerebellum plays a prominent role in mentalization and a moderate role in mirroring, and these social cognitive functions were mostly located in the posterior lobules spanning both the vermis and lateral hemispheres (Van Overwalle et al. 2014; Guell et al. 2018). Consistent with this, limited evidence suggests that loneliness is linked with cerebellar functions. Higher loneliness levels were associated with increased resting-state functional connectivity between the cerebellum and the postcentral gyrus and insula, even when objective social isolation was controlled for (Layden et al. 2017). While evidence on cerebellar functional activations in associate with loneliness is lacking, a body of literature suggests that the level of cerebellar activities during affective processing, such as during performing the emotional stroop task, is altered in major depressive individuals (Fitzgerald et al. 2008; Mitterschiffthaler et al. 2008; Groenewold et al. 2013).

The current study aimed to explicitly assess the association between the structure and functions of the cerebellum, and individual difference in susceptibility to perceiving loneliness, measured by taking into account both self-reported loneliness levels and objective social connections. We hypothesised that (1) alterations in cerebellar volume, particularly in the vermis and in the posterior lobules, would be observed in participants reporting greater susceptibility to social isolation. To further elucidate the implications of the affect processing and regulatory functions of the cerebellum in susceptibility to social isolation, we also examined cerebellar functional MRI activations, while participants were performing an emotion-word Stroop task. This task recruits both higher level cognitive control networks involved in response selection and executive functions, and ‘emotion networks’ implicated in affect processing and regulation (Mitterschiffthaler et al. 2008). Based on the limited existing literature, we also tentatively hypothesised that (2) the vermis and the posterior lobules of the cerebellum would exhibit altered activities and increased functional connectivity with the prefrontal cortex in relation to loneliness, particularly among those showing higher levels of susceptibility to social isolation.

The current study specifically examined the volumetric and functional indices of the cerebellum in relation to perceived loneliness, objective social network, and susceptibility to isolation. We found that compared to individuals reporting robust (perceiving lower level of loneliness despite relatively high social isolation) or concordant (reporting similar levels of loneliness and social isolation) pattern of susceptibility to isolation, those who were particularly vulnerable to perceiving loneliness despite relatively low social isolation exhibited increased gray matter volume of the posterior cerebellum encompassing the Crus I of the vermis and the lobule IX. Specifically during processing positive stimuli in the emotion stroop task, the association between loneliness and the activities in the right posterior cerebellum was more negative in the concordant individuals relative to both robust and susceptible individuals, while the association between social network and activities in the right posterior cerebellum was more positive among concordant individuals relative to the other two groups. Finally, during positive processing, loneliness positively predicted the right posterior cerebellum functional connectivity with the right visual cortex, and social network positively predicted the right posterior cerebellum functional connectivity with the right premotor cortex. Collectively, these findings provide novel evidence on the intricate role of the cerebellum in underlying not only the experience of loneliness, but also the susceptibility to perceiving loneliness.

A large body of clinical work indicates that damage to the posterior cerebellum is associated with cognitive deficits in executive control (e.g., planning, set-shifting, working memory), whereas vermis lesions are commonly linked to behavioural-affective disturbances (Baillieux et al. 2008; D’Angelo and Casali 2013). According to a meta-analysis, the cerebellar areas commonly activated in social cognitive tasks such as mentalization and mirroring were primarily located in the medial and posterior divisions, which also overlap extensively with the areas activated in executive control tasks (Van Overwalle et al. 2014). One possibility is that the cerebellum may not perform a domain-specific role in social cognition, but rather provides general cognitive support for complicated social processes (Van Overwalle et al. 2014). Recent theories on cerebellar functions posit that the cerebellum utilizes a ‘forward model’ based on stored associations, which uses feedback information to compute predictions and errors, which are in turn forwarded to the prefrontal and associative cortices (D’Angelo and Casali 2013). Such functions support the role of the cerebellum in prioritizing, processing and coordinating behaviours in response to incoming information, and in general cognitive and emotion control (D’Angelo and Casali 2013). Individuals with major depressive disorder, who typically exhibit deficits in both cognitive control and emotion regulation (Fitzgerald et al. 2008; Vasic et al. 2009), were found to show reduced volumes of the vermis and cerebellar atrophy (Beyer and Krishnan 2002). However, the susceptible group in the current study exhibited the opposite pattern of increased structural volumes of the posterior and medial cerebellum. Given these individuals were psychologically healthy, it could be that the increased cerebellar volume was an adaptive mechanism which compensates for the possible deficiencies in social, affective and/or cognitive functions that underlie those individuals’ high tendency to perceive loneliness despite relatively intact social network. Our current findings could not provide definitive evidence on which neural mechanisms might underlie those deficiencies, but one possibility is reduced cerebellar functional efficiency that could be related to its altered functional connectivity patterns. From a dispositional point of view, those susceptible individuals may be low on social reward dependence, a trait characterized by high sensitivity to social rewards and positive attitudes towards social relationships (Lebreton et al. 2009). Indeed, bilateral cerebellar gray matter was found to show inverse relationship with social reward dependence in the general population, which was considered to be potentially linked to the role of the cerebellum in signaling anticipation of aversive events (Ploghaus et al. 2003; Wiech et al. 2008; Lebreton et al. 2009). Thus, the increased cerebellar structural volume in loneliness-susceptible individuals may be considered as a general consequence of increased cerebellar involvement in socioaffective and cognitive processes. We also observed a negative association of age and medial cerebellar volume. This association could be primarily due to the protracted decrease of cerebellar volume which extends beyond young adulthood (Tiemeier et al. 2010; Wierenga et al. 2014). Nevertheless, the between-group difference in cerebellar volume was independent of and not moderated by the age effect.

The three participant groups performed comparably on the emotion stroop task, which might not be surprising given all participants were psychologically healthy with no neurological conditions. Thus, the observed between-group difference in brain activations during task performance was unlikely to be due to performance factors. Despite lack of main effect, we found intricate interactive effects of loneliness and susceptibility to isolation, and of social network and susceptibility to isolation. The right Crus I of lobule VII and the lobule IX were both associated with affective processing of positive and negative stimuli (Rohr et al. 2013; Keren-Happuch et al. 2014; Guell et al. 2018), whereas the Crus I and Crus II were both linked with executive functions (Van Overwalle et al. 2014; Guell et al. 2018). Our findings revealed effects of loneliness and susceptibility to isolation on posterior and medial cerebellar functions specifically in the positive word condition, highlighting the important role of hedonic processing and reward sensitivity in loneliness. Interesting, a previous study reported that patients with cerebellar stroke showed reduced pleasant experience but unchanged unpleasant experience (Turner et al. 2007), confirming the notion that the cerebellum performs essential hedonic functions. Loneliness showed particularly negative relationship with the right posterior and medial cerebellum activity to positive word stimuli in concordant participants, suggesting that for these individuals, the level of loneliness experience is primarily negatively determined by the person’s hedonic responsiveness. Notably, the cerebellum is also heavily implicated in anticipatory and attention functions (Ploghaus et al. 2003; D’Angelo and Casali 2013). Based on joint consideration of the socioaffective and cognitive processing that take place in the posterior and medial cerebellum, we speculate that individuals with increased cerebellar activity are more capable of engaging in anticipatory and consummatory hedonic processing during social contexts, which would in turn prompt those individuals to seek further social interactions, with the consequence of reduced loneliness levels. Conversely, individuals with decreased cerebellar activity might show reduced hedonic processing during social interactions, which results in decreased positive social expectation and/or processing, leading to avoidance of social interactions and elevated loneliness feeling. These explain the high contingency of loneliness feeling on social network size for the concordant participants. On the other hand, the association of loneliness and cerebellar activities to positive words was less marked and less negative in the susceptible and robust participants. It could be that the cerebellar activities in robust participants were more stable in affective responses, which might be due to greater affective regulatory and reappraisal functions in the anterior cingulate and lateral prefrontal cortices (Mohanty et al. 2007; Wiech et al. 2008; Kohn et al. 2014). As a result, the affective states and loneliness feelings of those individuals may be less reliant on the size of their social network. While we observed no significant between-group difference in prefrontal or cingulate activations during task performance, future studies that employ more socioaffectively salient tasks might shed more light on the neural functional characteristics of robust individuals. The lack of negative association between cerebellar activities to positive stimuli and loneliness in susceptible individuals might reflect low efficiency of hedonic processing in the cerebellum, and/or reduced negative affect regulatory functions (Wiech et al. 2008; D’Angelo and Casali 2013). Furthermore, the right posterior cerebellar activities to positive words were positively associated with social network size in concordant participants. Given increased social network typically entail richer social interactions and greater anticipation of social rewards, the increased cerebellar responsiveness to hedonic stimuli may reflect the individual’s generally high reward sensitivity when social support is satisfactory (Grippo et al. 2007). Such positive association was absent in susceptible and robust participants, further highlighting the potential deficits in social reward processing and responsiveness in the former group, and high stability of affect processing systems in the latter group. It should be emphasized that due to the cross-sectional nature of the current study and the lack of tasks that overtly assess social emotions, the interpretation of our findings necessarily remain speculative and should be considered with caution.

Main effects of loneliness and social network size on cerebellum functional connectivity during processing positive words in the emotion stroop task were observed, which did not differ across the susceptibility groups. Specifically, loneliness predicted more positive connectivity between the right lobule IX and the visual cortex. A meta-analysis suggested that the right lobule IX is commonly activated in emotion processing tasks, including the emotion stroop (Keren-Happuch et al. 2014). Another meta-analysis suggests that the right lobule IX is implicated in mentalizing processes (Van Overwalle et al. 2014). Moreover, recent evidence supports that the lobule IX performs both social and language functions (Guell et al. 2018). Both the posterior cerebellum and the occipital cortex were found to be activated during various executive control processes such as inhibition, conflict processing and attention (Chantiluke et al. 2012; Niendam et al. 2012), and occipital activations during performing the stroop task are commonly reported (Banich et al. 2000). A recent study reported that among individuals with autism spectrum disorder characterized by deficient social cognitive functions, resting-state functional connectivity between the cerebellum and visual areas was increased (Oldehinkel et al. 2019). Such increased connectivity may underlie increased cognitive and perceptual efforts in processing incoming social information for individuals with decreased social cognitive functions, which could be either the cause or consequence of loneliness (Hawkley and Cacioppo 2010). The premotor cortex is involved in both cognitive (e.g., attention, inhibition) and motor control functions (Abe and Hanakawa 2009). The premotor cortex also contains large populations of ‘mirror neurons’ that fire when observing other people’s actions (Heyes 2010), which are critical for inferring intention based on self-reflection. Interestingly, the posterior cerebellum is also heavily involved in mirroring processes (Van Overwalle et al. 2014), and the resting-state functional connectivity between the right cerebellum and the premotor cortex was found to be reduced in autistic children (Verly et al. 2014). Increased social network size would provide greater opportunities to observe and interact with others, which allows the individual to engage in more mirroring and theory of mind processes. Notably, both the effects of loneliness and social network size on cerebellar connectivity were independent of the susceptibility group status, indicating that the connectivity patterns were associated with actual social support and loneliness experience rather than with individual difference in proneness to perceiving loneliness. Also, both effects were observed on connectivity during processing positive words, suggesting that the brain hedonic processes, which affects the individual’s sensitivity to social rewards, may be key associates of loneliness and social support. An independent positive association between age and the cerebellar-premotor cortex connectivity was also revealed, a result that is consistent with previous literature on the development of fronto-cerebellar circuitries (Rubia et al. 2007). The implication of this age-associated change in key cerebellar connectivity remains unclear and demands future investigation.

Several limitations of the present study can be addressed in future research. We did not include tasks that explicitly assess social cognition or emotion, thus our interpretation on the role of the cerebellum in social processing remained tentative. Future studies may include mentalization, theory of mind or social emotion tasks. Second, we did not explicitly assess the participants’ state or trait affect, which could have affected the results. Future studies may additionally control for both state and trait positive and negative affects. Third, trait constructs such as empathy and social reward dependence could be measured in future studies to elucidate their relationship with susceptibility to perceiving loneliness. Finally, our participant sample encompassed a large age range, which increased the generalizability of the findings but also the heterogeneity of the participants in neural and socioaffective development. Future research should validate our findings on samples with more restricted ages.

In this study, we made novel discovery on the intricate role of the posterior and medial cerebellum in underlying an individual’s proneness to perceiving loneliness, and how such proneness may further interact with loneliness and social network size in determining the cerebellar responses to positive stimuli. Our findings lay the ground for future research to more finely delineate the role of the cerebellum in loneliness, and suggests that the social cognitive and affective processing of positive stimuli may be key associate of loneliness proneness. Our findings have important implications for interventions targeted at reducing loneliness levels in the general population, which need to take into account individual differences in susceptibility to social isolation, for which the cerebellum is a putative key neural correlate.