The present study investigates the potential relationship between changes in perceived stress and morphological changes in the amygdala. As predicted, there was a significant correlation between changes in PSS scores and changes in amygdaloid gray matter density. The more participants’ stress levels decreased, the greater the decrease of gray matter density in the right amygdala.
The amygdala is widely regarded as one of the most important limbic structures in prevailing models of stress states and anxiety disorders. It receives information from sensory modalities and projects to other subcortical structures, thereby mediating stress-related behavioral and physiological effects such as stress-hormone release, blood pressure elevation and facial expression of fear (LeDoux, 2000
). The cluster identified here appears to be located in the lateral/basolateral region of the amygdala (Mai et al.
). The basolateral region has been proposed to serve as the site for the relay of sensory information from subcortical and cortical sensory areas to the central nucleus of the amygdala during anxiety responses (Campeau and Davis, 1995
). Evidence of stress-related plasticity in these regions has previously been found in animal studies, including increased dendritic length and arborization within the basolateral complex of the amygdala (Vyas et al.
; Mitra et al.
). Strikingly, the basolateral amygdaloid sub-region identified in these rodent studies corresponds to the region identified here. Cytoarchitectural modifications such as those observed in rodent studies could potentially contribute to the increased gray matter density observed in a subset of the individuals in the present study. However, studies designed to establish the cellular mechanisms underlying the observed differences in amygdaloid gray matter in humans would require postmortem investigations.
Our results indicated an association between changes in stress levels and morphometric changes in the right, but not the left amygdala. It has been suggested that the right amygdala mediates an initial, fast and perhaps automatic stimulus detection, followed by a more evaluative and discriminative response by the left amygdala (Morris et al.
; Wright et al.
; Glascher and Adolphs, 2003
; Costafreda et al.
). Based on this model, our data suggest that this stress reduction intervention may strongly impact the participants’ initial reaction to stimuli. This is consistent with a recent study demonstrating decreased autonomic arousal (skin conductance response) to affective stimuli following a stress reduction course similar to the one in this study (Ortner et al.
). However, further research will be required to directly test any relationship between gray matter changes and reactions to stimuli.
Previous longitudinal structural MRI studies in humans have shown that repeated activation of a neural region, either while learning new skills (Draganski et al.
; Ilg et al.
) or through transcranial magnetic stimulation (May et al.
), leads to an increase in the corresponding regional gray matter, whereas cessation of activation leads to a decrease. It seems plausible that this pattern could apply to the present findings—that changes in stress facilitate changes in amygdala activity, which in turn mediate changes in gray matter density. Interestingly, in rats, removal of experimental stressors after a period of chronic exposure did not lead to a reversal of the identified amygdaloid neuronal hypertrophy, or to the reversal of the associated enhanced anxiety-like behaviors within the observed time-frame of 21 days (Vyas et al.
). Our results suggest that ameliorating the subjective experience of stress through a behavioral intervention may actually decrease amygdala gray matter density in humans. This finding is particularly interesting as it suggests that an active re-learning of emotional responses to stress (such as taught in MBSR) can lead to beneficial changes in neural structure and well-being even when there is presumably no change in the person’s external environment. Future research will be required to address whether stress-induced alterations in the basolateral complex of the amygdala might influence a person’s susceptibility to anxiety and other affective disorders (Sajdyk et al.
; Shekhar et al.
Gianaros et al.
) recently reported that lower gray matter volume in the bilateral amygdala predicted greater stressor-related amygdala activation, as well as greater blood pressure reactivity. However, the complexity and heterogeneity of amygdala subnuclei, in addition to the low spatial resolution of neuroimaging methods, make interpreting this seemingly contradictory finding difficult. As methods and technology improve, future studies could consider how effects of stress may vary across the several heterogeneous subregions of the amygdala. It should also be noted that Gianaros et al.
) investigated gray matter volume, which is distinct from gray matter density examined in the present study. The biological differences underlying these two neuroimaging techniques remain unclear.
Although a correlation was found between changes in amygdaloid structure and perceived stress, the present study did not show a significant overall main effect of the training on amygdaloid gray matter density. Thus, the results do not support the conclusion that MBSR training per se leads to decreases in gray matter in this region. As reported elsewhere (Hölzel et al., under review), main effect analyses on a sub-cohort of the study participants did reveal significant changes in hippocampal, inferior temporal lobe, posterior cigulate, temporo-parietal and cerebellar gray matter density, though these regions were not correlated with changes in perceived stress.
The scatter plot (D) illustrates that amygdaloid gray matter density increased for some participants, though it should be noted that a lot of those subjects also reported increases in perceived stress following the MBSR program. Some of the participants with improved perceived stress scores appear to have slight increases in gray matter density, but these small deviances may reflect noise. Alternatively, changes in amygdala gray matter may be temporally delayed relative to changes in perceived stress, perhaps requiring habitual activation in this region to subside prior to longer term structural changes. The results do support a bidirectional correlation; further work will be required to determine the precise relationship between the self-report measure and cellular changes. PSS values and gray matter density were not correlated at the pre-intervention time-point. This is in line with previous findings (Gianaros et al.
) and is not unexpected, as numerous factors can influence brain gray matter variables (Meyer-Lindenberg et al.
; Wilke et al.
). Importantly, we assessed the relationship between the change in one variable, namely perceived stress and changes in gray matter density within the amygdala. By employing a longitudinal design most within-subject variables were kept relatively constant, while the factor of interest, perceived stress, varied. Some behavioral variables, such as smoking, diet or exercise, and psychological factors (e.g. neuroticism) can also co-vary with changes in perceived stress, however, and might mediate or drive the relationship between changes in perceived stress and structural changes (cf., Gianaros et al.
). These variables were not assessed in the current study, and so it is unknown if the relationship between perceived stress and gray matter observed here is direct or indirect.
Several previous cross-sectional studies have investigated the impact of mindfulness meditation on brain morphology by comparing groups of experienced mindfulness meditators to nonmeditators (Lazar et al.
; Pagnoni and Cekic, 2007
; Hölzel et al.
; Luders et al.
). These studies identified several regions of altered brain morphology, but none within the amygdalae. However, none of these studies assessed the participants’ perceived stress levels. Again, these data highlight the limitations of the cross-sectional study design. The unique hypothesis-driven, focused analysis employed in the present study revealed a novel link between changes in amygdaloid gray matter density and decreases in self-reported stress following stress-reduction training, marking a significant advance in our understanding of the association between both. Whereas previous studies have demonstrated that gray matter modifications can result from the acquisition of abstract information (Draganski et al.
), motor skills (Draganski et al.
) and language skills (Mechelli et al.
), this is the first study to demonstrate neuroplastic changes associated with changes in a measure of a psychological state.
Conflict of interest