This is the first study to investigate effects of a mindfulness intervention on FC in resting-state networks in a preadolescent cohort with sub-clinical anxiety-related attention impairments. Understanding mindfulness-induced neural and associated behavioral changes is crucial to support the efficacy of early interventions (Sylvester et al., 2012). Our results showed increased FC in the resting-state networks investigated [SLN, FPN, and DMN (MTG)] following the mindfulness intervention. Increased FC was observed in the SLN and FPN, suggesting improvements in cognitive control and emotion regulation, however increased FC in the DMN in contrast to decreased FC in the DMN observed in adult studies may be associated with participants’ maturing brain networks (Brewer et al., 2011; Hasenkamp, Wilson-Mendenhall, Duncan, & Barsalou, 2012; Marusak et al., 2018). These observed changes in FC were associated with improved behavioral outcomes. Positive correlations were identified pre-intervention in the FPN to several brain regions for the HRT and the SCAS total and in the DMN to several brain regions for the HRT.
The SLN, including key nodes insula and ACC (Uddin, 2015), is involved in the detection, filtering and orientation to salient stimuli (Vinod Menon & Uddin, 2010). Decreased FC in the SLN to the prefrontal regions and insula has been implicated in deficits in emotion regulation and cognitive control, processes recognized as underlying anxiety and attention disorders (Etkin, Egner, & Kalisch, 2011; Uddin, Supekar, Ryali, & Menon, 2011).
Recent research has associated mindfulness interventions with increased FC in the SLN to the insula, vmPFC, dmPFC, dlPFC, ACC, hippocampus and amygdala and improvements in anxiety symptoms (Marusak et al., 2018; Tang et al., 2015), consistent with our findings. Specifically, increased FC in the prefrontal regions has been associated with improved cognitive control required for effective processing of emotionally salient stimuli (Etkin et al., 2011; Kober et al., 2008). Furthermore, our observation of increased SLN connectivity to the dlPFC, hippocampus and amygdala has been associated with improvements in switching attention from emotionally salient events. (Forster, Nunez Elizalde, Castle, & Bishop, 2015; Zheng et al., 2017).
The FPN is the central executive network, and with frontal regions, supports executive control of attention (Corbetta & Shulman, 2002; Petersen & Posner, 2012). Decreased FC in the FPN is associated with deficits in executive function, specifically cognitive control and emotion regulation and has been reported in children with impairments in attention and anxiety (Rosenberg et al., 2016). Mindfulness training has been associated with increased FC in the FPN and increased cognitive control and emotion regulation in adult and youth populations (Hölzel et al., 2013; Vago & David, 2012). Our findings indicate increased FC to the vmPFC, dlPFC, vmPFC, IFG, MFG, and insula in our cohort. The insula has been implicated in introspective thoughts (Critchley, Wiens, Rotshtein, Öhman, & Dolan, 2004), thus increased FC to the dlPFC and vmPFC may modulate attention to anxious thoughts (Vinod Menon & Uddin, 2010; Namkung, Kim, & Sawa, 2017). Moreover, an overall decrease in participants’ anxiety scores was supported by improved FC in these regions. Increased FC to the IFG and MFG has been implicated in increased response inhibition to irrelevant internal emotional stimuli, thus supporting sustained attention (Japee, Holiday, Satyshur, Mukai, & Ungerleider, 2015; Petersen & Posner, 2012). Furthermore, an overall decrease in participants’ commission scores supports improvement in inhibitory control.
Attention to self-referential thoughts or ‘mind-wandering’ has been associated with increased activation in the DMN during rsfMRI, as well as anxiety and attention issues (Brewer et al., 2011; Marusak et al., 2018). Mindfulness training has been shown to decrease activation in the DMN in adults, thus decreasing mind-wandering (Jha et al., 2007). Our result of increased FC within the DMN regions of dmPFC, ACC, IFG, MFG, and SFG may underpin improvements in emotion regulation and corroborate recent findings by (Hafeman et al., 2020). However, our findings are in contrast with another study’s findings of decreased FC between the DMN and dlPFC (Bauer et al., 2020). Previous studies established that developmental FC changes in children are characterized by weakening short-range FC and strengthening long-range FC (Kelly et al., 2009; Uddin, Supekar, & Menon, 2010). Our findings of increased FC were observed after a relatively short period of 10-12 weeks between MRI scans. Taken together, we suggest that FC changes within networks in this study are due to the mindfulness intervention, and less likely associated with developmental changes. Furthermore, increased FC between the SLN, FPN and DMN following the mindfulness intervention may indicate improved cognitive control and emotion regulation (Marusak et al., 2018). The SLN plays a crucial role in switching activation between the FPN and DMN, as these networks support processing the emotional salience of irrelevant stimuli and are often dysfunctional in children with anxiety-related attention impairments (Sha et al., 2019).
At pre-intervention, the right FPN was correlated with slower response times, often associated with inattentiveness. Specifically, increased FC to the left amygdala and right hippocampus sends direct projections to the caudate, negatively affecting goal-directed behavior (Grahn, Parkinson, & Owen, 2008). The FPN connectivity to the caudate showed a positive correlation with SCAS total score, indicating increased FC is associated with elevated anxiety symptoms. Correspondingly, greater caudate activation has been reported in children with high levels of behavioral inhibition, considered a risk factor for anxiety (Lahat, Benson, Pine, Fox, & Ernst, 2018; Vincent, Kahn, Snyder, Raichle, & Buckner, 2008). However, this response was improved in our participants following the mindfulness intervention, with no significant correlations observed between the FPN FC and the HRT or total SCAS score, suggesting normalization of participants’ FC.
Momentary lapses of attention have been associated with increased activity in the DMN wherein strong interactivity of key nodes PCC/precuneus, may be linked to slower HRT (Cavanna & Trimble, 2006; Weissman, Roberts, Visscher, & Woldorff, 2006). Moreover, the correlation of DMN connectivity correlation with HRT might arise from increased self-awareness and reduced task performance, which is known to take place in participants with behavioral deficits (Alexopoulos et al., 2012; Zhang & Li, 2012). At post-intervention, these observed correlations of the DMN connectivity were not significant, suggesting the mindfulness intervention may have improved FC and behavioral performance (Brewer et al., 2011).
Taken together, the lack of correlations post-intervention in the FPN and DMN networks suggests the mindfulness intervention altered FC in these networks and reduced behavioral scores. Thus, the observed pre-intervention correlations and the behavioral variables may be markers for anxiety-related attention impairments in pre-adolescent populations.
Our study has several limitations that require mentioning. Firstly, a control group was not included in the design, as sub-clinical anxiety is pervasive in preadolescent children, with approximately up to 70% of children reporting they worry “every now and then” (Cartwright-Hatton, 2006; Muris, Merckelbach, Meesters, & van den Brand, 2002). Secondly a mindfulness scale was not included, as the concepts of metacognition and interoceptive awareness are still developing during pre-adolescence; thus, researcher involvement to clarify questions may have increased risk of response bias. Finally, as this was a group program, some improvements may be attributed to social interactions and group discussions.