Neuroanatomic differences associated with stress susceptibility and resilience

In chronically stressed mice, social avoidance z score was negatively correlated with peak voxel volume in the cingulate cortex (r = −.79, p < .001).

“In addition, we found a significant negative correlation between social avoidance scores and peak voxel volumes of the cingulate cortex (r = −.79, p < .001) (Figure 2B), suggesting that the cingulate cortex is smaller in stress-susceptible mice.”

• Also stated in: § DISCUSSION · plain

  “Our structural analysis revealed that gray matter volume of the VTA increased as a function of social avoidance, whereas cingulate cortex and NAc volumes declined with increasing social avoidance.”

In chronically stressed mice, social avoidance z score was positively correlated with peak voxel volume in the ventral tegmental area (r = .69, p < .001).

“Social avoidance scores correlated positively with VTA peak voxel volume (r = .69, p < .001) (Figure 2A), supporting a role of the VTA in promoting stress susceptibility.”

• Also stated in: § DISCUSSION · plain

  “Our structural analysis revealed that gray matter volume of the VTA increased as a function of social avoidance, whereas cingulate cortex and NAc volumes declined with increasing social avoidance.”

In control mice, hippocampus and cingulate cortex volumes were positively correlated (r = .6, p = .001).

“In control mice (n = 12), we observed correlations only between the hippocampus and hypothalamus (r = .57, p = .04) (Supplemental Figure S6C) and between the hippocampus and cingulate cortex (r = .6, p = .001) (Supplemental Figure S6E), but not between any of the other brain regions analyzed (Supplemental Figure S6).”

Hippocampal volume was positively correlated with VTA volume (r = .64, p = .003).

“Hippocampal volume is positively correlated with VTA volume (r = .64, p = .003) (Figure 4D) and negatively correlated with cingulate cortex volume (r = −.68, p = .001) (Figure 4E), suggesting an interaction of the hippocampus with the VTA and medial PFC (4).”

In the VTA, NAc, and cingulate cortex, fractional anisotropy correlated with behavioral z scores in the opposite direction from anatomic volume.

“Our DTI analysis revealed that FA correlated with behavioral z scores in opposite directions than anatomic volume in the VTA, NAc, and cingulate cortex.”

In chronically stressed mice, social avoidance z score was positively correlated with peak voxel volume in the hypothalamus (r = .77, p < .01).

“We found significant positive correlations between social avoidance scores and peak voxel volumes of the hippocampus (r = .81, p < .001) (Figure 3A) and hypothalamus (r = .77, p < .01) (Figure 3B), suggesting that larger hypothalamus and hippocampal volumes are associated with stress susceptibility.”

DTI analysis showed a positive correlation between behavioral z scores and fractional anisotropy in the hippocampus.

“In addition, our DTI analysis revealed positive correlations between behavioral z scores and FA in the hippocampus and hypothalamus, indicating that diffusion in these structures is more anisotropic in susceptible mice (Figure 3A, B; graphs on the right).”

There was a trend toward a small negative correlation between VTA and NAc volumes (r = −.14, p = .06).

“We observed a trend toward a small negative correlation between the VTA and NAc in our analysis (r = −.14, p = .06) (Figure 4B).”

Atlas-based comparisons confirmed that susceptible mice had larger hippocampal volume than resilient mice.

“Atlas-based comparisons of group-level structural differences confirmed larger hippocampal volume in susceptible (n = 11) compared with resilient (n = 8) mice (Supplemental Table S1 and Figure S4).”

Hippocampal volume was negatively correlated with cingulate cortex volume (r = −.68, p = .001).

“Hippocampal volume is positively correlated with VTA volume (r = .64, p = .003) (Figure 4D) and negatively correlated with cingulate cortex volume (r = −.68, p = .001) (Figure 4E), suggesting an interaction of the hippocampus with the VTA and medial PFC (4).”

In control mice, no significant correlations were found between social avoidance z score and volume of the same brain structures.

“No significant correlations were observed between social avoidance z score and volume of the same brain structures in control mice (n = 12) (Supplemental Figure S5).”

Peak voxel volumes of the VTA and cingulate cortex were negatively correlated (r = −.61, p = .01).

“We found a negative correlation between peak voxel volumes of the VTA and the cingulate cortex (r = −.61, p = .01) (Figure 4A).”

In chronically stressed mice, social avoidance z score was positively correlated with peak voxel volume in the amygdala (r = .77, p < .01).

“Social avoidance scores also correlated negatively with peak voxel volumes of the NAc (r = −.48, p < .05) (Figure 2C) and positively with peak voxel volumes of the amygdala (r = .77, p < .01) (Figure 2D) (25–27).”

In control mice, hippocampus and hypothalamus volumes were positively correlated (r = .57, p = .04).

“In control mice (n = 12), we observed correlations only between the hippocampus and hypothalamus (r = .57, p = .04) (Supplemental Figure S6C) and between the hippocampus and cingulate cortex (r = .6, p = .001) (Supplemental Figure S6E), but not between any of the other brain regions analyzed (Supplemental Figure S6).”

There was a trend toward a negative correlation between hippocampus and NAc volumes (r = −.44, p = .06).

“We found a trend toward a negative correlation between the volume of the hippocampus and NAc (r = −.44, p = .06) (Figure 4F).”

Social avoidance z score was positively correlated with ventral hippocampus volume (r = .81, p < .001).

“We found similar correlations when separately analyzing the dorsal (r = .80, p < .001) and ventral (r = .81, p < .001) hippocampus (Supplemental Figure S3).”

Susceptible and resilient mice clustered at opposite ends of the VTA-cingulate cortex volume correlation.

“Susceptible and resilient mice cluster at opposite ends of this correlation (Figure 4A; open circles, susceptible mice, n = 11; solid circles, resilient mice, n = 8).”

An interaction model found no significant group-dependent interactions for structure-structure correlations.

“Using this analysis, we did not find any interactions that were significant (Supplemental Table S3).”

Susceptible and resilient mice clustered at opposite ends of the hippocampus-hypothalamus volume correlation.

“Similar to our findings for the VTA and cingulate cortex, susceptible (n = 11, open circles) and resilient (n = 8, solid circles) mice cluster at opposite ends of this correlation, suggesting that stress susceptibility is associated with stronger structural covariance between the hippocampus and hypothalamus.”

Social avoidance z score was positively correlated with dorsal hippocampus volume (r = .80, p < .001).

“We found similar correlations when separately analyzing the dorsal (r = .80, p < .001) and ventral (r = .81, p < .001) hippocampus (Supplemental Figure S3).”

In the amygdala, fractional anisotropy-behavior and volume-behavior correlations had the same direction.

“In the amygdala, FA-behavior and volume-behavior correlations had the same direction (Supplemental Figure S2).”

In chronically stressed mice, social avoidance z score was negatively correlated with peak voxel volume in the nucleus accumbens (r = −.48, p < .05).

“Social avoidance scores also correlated negatively with peak voxel volumes of the NAc (r = −.48, p < .05) (Figure 2C) and positively with peak voxel volumes of the amygdala (r = .77, p < .01) (Figure 2D) (25–27).”

• Also stated in: § DISCUSSION · plain

  “Our structural analysis revealed that gray matter volume of the VTA increased as a function of social avoidance, whereas cingulate cortex and NAc volumes declined with increasing social avoidance.”

In chronically stressed mice, social avoidance z score was positively correlated with peak voxel volume in the hippocampus (r = .81, p < .001).

“We found significant positive correlations between social avoidance scores and peak voxel volumes of the hippocampus (r = .81, p < .001) (Figure 3A) and hypothalamus (r = .77, p < .01) (Figure 3B), suggesting that larger hypothalamus and hippocampal volumes are associated with stress susceptibility.”

Across all chronically defeated mice, hippocampal volume and hypothalamus volume were positively correlated (r = .62, p = .005; n = 19).

“To investigate whether hippocampus-hypothalamus connectivity may determine individual differences in stress sensitivity, we correlated hippocampal volume with hypothalamus volume and found a positive correlation across all chronically defeated mice (r = .62, p = .005; n = 19) (Figure 4C).”

The larger hippocampal volume associated with social avoidance was specifically observed in CA3.

“Larger hippocampal volume was specifically observed in CA3 (red areas in Figure 3A), a region of the hippocampus that is particularly vulnerable to the effects of stress (8,31).”

DTI analysis showed a positive correlation between behavioral z scores and fractional anisotropy in the hypothalamus.

“In addition, our DTI analysis revealed positive correlations between behavioral z scores and FA in the hippocampus and hypothalamus, indicating that diffusion in these structures is more anisotropic in susceptible mice (Figure 3A, B; graphs on the right).”

Social avoidance scores were positively correlated with cerebellum volume (r = .78, p < .001).

“the periaqueductal gray (r = .66, p = .002) (Figure 5F), a region involved in fear responses and defensive behavior (43,44); and the cerebellum (r = .78, p < .001) (Figure 5G), which has been implicated in neuropsychiatric disorders (45).”

In control mice, no correlations were found between the other analyzed brain regions.

“In control mice (n = 12), we observed correlations only between the hippocampus and hypothalamus (r = .57, p = .04) (Supplemental Figure S6C) and between the hippocampus and cingulate cortex (r = .6, p = .001) (Supplemental Figure S6E), but not between any of the other brain regions analyzed (Supplemental Figure S6).”

Social avoidance scores were negatively correlated with raphe nuclei volume (r = −.64, p = .003).

“the raphe nuclei (r = −.64, p = .003) (Figure 5B), which are importantly involved in regulating serotoninergic tone (40,41);”

Atlas-based comparisons confirmed smaller corpus callosum volume in susceptible mice compared with resilient mice.

“This additional analysis confirmed smaller BNST and corpus callosum volume in susceptible compared with resilient mice (Supplemental Table S2 and Figure S4).”

In control mice, no significant correlations were found between social avoidance and volume or fractional anisotropy of the novel brain regions.

“We did not find any significant correlations between social avoidance and volume or FA of any of these brain regions in control mice (n = 12).”

Social avoidance scores were positively correlated with periaqueductal gray volume (r = .66, p = .002).

“the periaqueductal gray (r = .66, p = .002) (Figure 5F), a region involved in fear responses and defensive behavior (43,44); and the cerebellum (r = .78, p < .001) (Figure 5G), which has been implicated in neuropsychiatric disorders (45).”

For the BNST, thalamus, habenular commissure, periaqueductal gray, and cerebellum, fractional anisotropy-behavior correlations mostly occurred in the opposite direction from volume-behavior correlations.

“Similar to our observation in the VTA, cingulate cortex, and NAc, most of these correlations occurred in opposite directions than anatomic volume, again suggesting that increased volume may be associated with less anisotropic diffusion in these brain regions.”

DTI analysis showed correlations between fractional anisotropy and behavioral z scores for the BNST, thalamus, habenular commissure, periaqueductal gray, and cerebellum.

“Our DTI analysis showed correlations in FA with behavioral z scores for the BNST, thalamus, habenular commissure, periaqueductal gray, and cerebellum (Supplemental Figure S7).”

Social avoidance scores were negatively correlated with thalamus volume (r = −.79, p < .001).

“the thalamus (r = −.79, p < .001) (Figure 5C), which controls consciousness and information relay; and the corpus callosum (r = −.77, p < .001) (Figure 5D), the central white matter tract connecting both cerebral hemispheres.”

Social avoidance scores were negatively correlated with BNST volume (r = −.78, p < .001).

“Social avoidance scores correlated negatively with the volume of the bed nucleus of the stria terminals (BNST) (r = −.78, p < .001) (Figure 5A), a brain structure involved in neuroendocrine regulation and anxiety-like behavior (36–39);”

Social avoidance scores were positively correlated with habenular commissure volume (r = .69, p = .001).

“Social avoidance scores also correlated positively with volume of the habenular commissure (r = .69, p = .001) (Figure 5E), a brain structure associated with depression and antidepressant response (42);”

Social avoidance scores were negatively correlated with corpus callosum volume (r = −.77, p < .001).

“the thalamus (r = −.79, p < .001) (Figure 5C), which controls consciousness and information relay; and the corpus callosum (r = −.77, p < .001) (Figure 5D), the central white matter tract connecting both cerebral hemispheres.”

Atlas-based comparisons confirmed smaller BNST volume in susceptible mice compared with resilient mice.

“This additional analysis confirmed smaller BNST and corpus callosum volume in susceptible compared with resilient mice (Supplemental Table S2 and Figure S4).”

Structural covariance between hippocampal CA3 and the hypothalamus was found in stressed and control mice.

“We found structural covariance between hippocampal CA3 and the hypothalamus in stressed as well as control mice, suggesting that structural synchrony between these regions generally exists in nonstressed individuals.”

In the study, local volume and FA correlated in opposite directions with social avoidance in the VTA, cingulate cortex, NAc, BNST, thalamus, periaqueductal gray, and cerebellum.

“In our study, local volume and FA correlate in opposite directions with social avoidance in the VTA, cingulate cortex, NAc, BNST, thalamus, periaqueductal gray, and cerebellum.”

Susceptible mice had reduced raphe nuclei volume.

“We also found that susceptible mice exhibit reduced volume of the raphe nuclei”

In the study, local volume and FA correlated positively with social avoidance in the hippocampus and hypothalamus.

“In contrast, local volume and FA positively correlate with social avoidance in the hippocampus and hypothalamus.”

Periaqueductal gray volume increased with stress susceptibility.

“We also found increased volume of the periaqueductal gray as a function of stress susceptibility”

The atlas-based differences were less significant than the voxelwise correlations.

“However, the differences were less significant”

Stress susceptibility is associated with stronger structural covariance between the hippocampus and hypothalamus.

“Similar to our findings for the VTA and cingulate cortex, susceptible (n = 11, open circles) and resilient (n = 8, solid circles) mice cluster at opposite ends of this correlation, suggesting that stress susceptibility is associated with stronger structural covariance between the hippocampus and hypothalamus.”

Chronic stress is a major risk factor for depression.

“Chronic stress is a major risk factor for depression (1); however, it is unclear why some individuals are more susceptible to the effects of stress than others.”

• cites [1] Kessler, 1997, The effects of stressful life events on depression, Annu Rev Psychol sourceNot checked yet

  Not checked in the free run.

Increased volume may be associated with less anisotropic diffusion in these brain regions.

“Similar to our observation in the VTA, cingulate cortex, and NAc, most of these correlations occurred in opposite directions than anatomic volume, again suggesting that increased volume may be associated with less anisotropic diffusion in these brain regions.”

Brain reward pathways, including the VTA and NAc, as well as the PFC and amygdala, have been implicated in the neural circuitry mediating individual differences in stress susceptibility.

“Brain reward pathways, including the ventral tegmental area (VTA) and the nucleus accumbens (NAc) (4) as well as the prefrontal cortex (PFC) and the amygdala (11,12), have been implicated in the neural circuitry mediating these differences in stress susceptibility.”

• cites [4] Chaudhury, 2013, Rapid regulation of depression-related behaviours by control of midbrain dopamine neurons, Nature sourceNot checked yet

  Not checked in the free run.

• cites [11] Narayanan, 2011, Social defeat: Impact on fear extinction and amygdala-prefrontal cortical theta synchrony in 5-HTT deficient mice, PLoS One sourceNot checked yet

  Not checked in the free run.

• cites [12] Fekete, 2009, Social defeat stress activates medial amygdala cells that express type 2 corticotropin-releasing factor receptor mRNA, Neuroscience sourceNot checked yet

  Not checked in the free run.

Chronic social defeat stress is a well-validated mouse model that consistently produces susceptible and resilient phenotypes.

“Chronic social defeat stress is a well-validated mouse model that consistently generates susceptible and resilient phenotypes (10).”

• cites [10] Krishnan, 2007, Molecular adaptations underlying susceptibility and resistance to social defeat in brain reward regions, Cell sourceNot checked yet

  Not checked in the free run.

Hippocampal connections to the hypothalamus inhibit glucocorticoid hormone release by suppressing the HPA axis.

“At the circuit level, hippocampal connections to the hypothalamus inhibit the release of glucocorticoid hormones via suppression of the hypothalamus-pituitary-adrenal (HPA) axis (13–15).”

• cites [13] Anacker, 2014, Early life adversity and the epigenetic programming of hypothalamic-pituitary-adrenal functionNot checked yet

  Not checked in the free run.

• cites [14] Anacker, 2011, The glucocorticoid receptor: Pivot of depression and of antidepressant treatment?, Psychoneuroendocrinology sourceNot checked yet

  Not checked in the free run.

• cites [15] Anacker, 2012, Can adult neurogenesis buffer stress responses and depressive behaviour?, Mol Psychiatry sourceNot checked yet

  Not checked in the free run.

The hippocampus governs HPA axis activity by modulating hypothalamic function.

“The hippocampus governs HPA axis activity by modulating the function of the hypothalamus (30).”

• cites [30] Surget, 2011, Antidepressants recruit new neurons to improve stress response regulation, Mol Psychiatry sourceNot checked yet

  Not checked in the free run.

The lack of a significant VTA-NAc correlation may suggest that structural differences in connectivity between the VTA, medial PFC, and NAc are primarily driven by covariance between the VTA and cingulate cortex.

“The lack of significant correlation between these two nuclei may suggest that, at the structural level, differences in connectivity between the VTA, medial PFC, and NAc may be primarily driven by covariance between the VTA and cingulate cortex.”

The absence of a correlation between social avoidance and brain volume in control animals suggests that different anatomy and connectivity may be a consequence of stress exposure.

“The fact that social avoidance in our study is not correlated with brain volume in control animals suggests that different anatomy and connectivity may be the consequence of stress exposure.”

After stress, susceptible mice show greater volume associations between the hippocampus and hypothalamus than resilient mice.

“However, after stress, susceptible mice show greater volume associations between these regions than resilient mice, as indicated by the finding that susceptible and resilient mice cluster at opposite ends of the correlation (Figure 3).”

Reduced BNST volume in susceptible mice may reflect network abnormalities that disturb hippocampal information relay to the hypothalamus.

“Reduced volume of the BNST in susceptible mice may reflect network abnormalities by which hippocampal information relay to the hypothalamus could be disturbed.”

• AssumptionRests on an assumption: Glutamatergic hippocampal outputs activate GABAergic inhibitory neurons of subcortical relay areas, including the BNST, which consequently inhibit neurons in the paraventricular nucleus of the hypothalamus.

Structural synchrony between hippocampal CA3 and the hypothalamus generally exists in nonstressed individuals.

“We found structural covariance between hippocampal CA3 and the hypothalamus in stressed as well as control mice, suggesting that structural synchrony between these regions generally exists in nonstressed individuals.”

Synchronized anatomic differences between the VTA and cingulate cortex may reflect altered connectivity between these regions in susceptible and resilient mice.

“These synchronized anatomic differences may reflect altered connectivity between these regions in susceptible and resilient mice.”

Disturbances in hippocampus-hypothalamus connectivity may be associated with enlarged hippocampal volume as compensatory changes aimed at reducing HPA axis hyperactivity.

“Such disturbances in hippocampus-hypothalamus connectivity may be associated with enlarged hippocampal volume as a result of compensatory changes in the hippocampus that are aimed at reducing HPA axis hyperactivity.”

• AssumptionRests on an assumption: Glutamatergic hippocampal outputs activate GABAergic inhibitory neurons of subcortical relay areas, including the BNST, which consequently inhibit neurons in the paraventricular nucleus of the hypothalamus.

Some studies have observed an association between the number of stressful life events and increased hippocampal volume in elderly, nondepressed subjects.

“Some studies have observed an association between the number of stressful life events and increased hippocampal volume in elderly, nondepressed subjects (53).”

• cites [53] Zannas, 2013, Negative life stress and longitudinal hippocampal volume changes in older adults with and without depression, J Psychiatr Res sourceNot checked yet

  Not checked in the free run.

Stress-integrative brain regions shape the neural architecture underlying individual differences in susceptibility and resilience to chronic stress.

“Stress-integrative brain regions shape the neural architecture underlying individual differences in susceptibility and resilience to chronic stress.”

The findings may at least in part reflect individual trait differences.

“Our findings, which confirm the result of this study, may at least in part also reflect individual trait differences.”

A recent longitudinal study suggested that stress-susceptible mice have larger hippocampal volume than resilient mice before stress onset.

“However, a recent longitudinal study suggested that stress-susceptible mice have larger hippocampal volume than resilient mice before stress onset (56).”

• cites [56] Tse, 2014, A longitudinal study of stress-induced hippocampal volume changes in mice that are susceptible or resilient to chronic social defeat, Hippocampus sourceNot checked yet

  Not checked in the free run.

Anatomic differences in the VTA, cingulate cortex, and NAc may be associated with their previously established functional role in mediating individual differences in stress susceptibility.

“These data support the notion that anatomic differences in these brain regions may be associated with their previously established functional role in mediating individual differences in stress susceptibility.”

Increased volumes of the hippocampus and hypothalamus may sensitize susceptible individuals to develop HPA axis abnormalities and social avoidance more than resilient mice.

“Increased volumes of two of the key regulators of the HPA axis may possibly sensitize susceptible individuals to develop HPA axis abnormalities and social avoidance more so than resilient mice.”

The hippocampus may belong to a broad neuroanatomic network tied to stress sensitivity.

“suggesting that the hippocampus may be part of a comprehensive neuroanatomic network associated with stress sensitivity”

The lower significance of the atlas-based differences indicates that the z score behavioral measure is important for distinguishing individuals.

“indicating the importance of the z score measure of behavior for differentiating between individuals.”

• AssumptionRests on an assumption: The z score measure of social avoidance is a valid and important behavioral measure for distinguishing susceptible and resilient individuals.

The increased periaqueductal gray volume suggests differences in fear and panic responses between susceptible and resilient mice.

“pointing toward differences in fear and panic responses between susceptible and resilient mice (43,44,59).”

• cites [43] Ballesteros, 2014, Effect of dorsal and ventral hippocampal lesions on contextual fear conditioning and unconditioned defensive behavior induced by electrical stimulation of the dorsal periaqueductal gray, PLoS One sourceNot checked yet

  Not checked in the free run.

• cites [44] Quintino-dos-Santos, 2014, Long-lasting marked inhibition of periaqueductal gray-evoked defensive behaviors in inescapably-shocked rats, Eur J Neurosci sourceNot checked yet

  Not checked in the free run.

• cites [59] Graeff, 2004, Serotonin, the periaqueductal gray and panic, Neurosci Biobehav Rev sourceNot checked yet

  Not checked in the free run.

Combining MRI and DTI data from the same brain regions may offer initial insight into tissue microstructure differences linked to structural differences between susceptible and resilient brains.

“Our multiparametric approach of combining MRI and DTI data from the same brain regions may provide some first insight into differences in tissue microstructure associated with structural differences between susceptible and resilient brains.”

The BNST projects to hypothalamic targets that produce fear responses.

“Moreover, the BNST sends outputs to hypothalamic targets to produce fear responses (38,39)”

• cites [38] Davis, 2010, Phasic vs sustained fear in rats and humans: Role of the extended amygdala in fear vs anxiety, Neuropsychopharmacology sourceNot checked yet

  Not checked in the free run.

• cites [39] Sink, 2011, Calcitonin gene-related peptide in the bed nucleus of the stria terminalis produces an anxiety-like pattern of behavior and increases neural activation in anxiety-related structures, J Neurosci sourceNot checked yet

  Not checked in the free run.

A simple gray-versus-white matter or cortical-versus-subcortical distinction does not appear sufficient to explain the observed behavior-FA and behavior-volume correlations.

“A simple distinction between gray and white matter or cortical and subcortical tissue does not seem sufficient to explain the observed behavior-FA and behavior-volume correlations.”

The systems-level analysis of anatomic volume differences between susceptible and resilient mice identified many brain regions potentially relevant to anxiety and depression.

“Our systems-level analysis of anatomic volume differences in susceptible and resilient mice revealed numerous brain regions with potential relevance to anxiety and depression.”

BNST projections to hypothalamic fear-response targets may be especially relevant to social avoidance caused by defeat.

“which may be particularly relevant for defeat-induced social avoidance.”

Brain structure, anisotropy, and cell number do not directly measure neuronal activity.

“In addition, a direct measure of neuronal activity is not provided by brain structure, anisotropy, or cell number.”

Considering the study and recent work, a consensus appears to be emerging linking structural alterations, tissue diffusivity, and underlying tissue microstructure in key regions such as the hippocampus.

“In light of our study and this recent work, a consensus seems to emerge between structural alterations, tissue diffusivity, and underlying tissue microstructure in key brain regions, such as the hippocampus.”

Microscopic changes in brain volume and anisotropy can arise from altered synaptogenesis, neuronal morphology, fiber density, and axonal branching, as well as from extraneuronal changes such as glial cell number or angiogenesis.

“Microscopic alterations in brain volume and anisotropy can be caused by changes in synaptogenesis, neuronal morphology, fiber density, and axonal branching as well as by extraneuronal changes, such as glial cell number or angiogenesis (65).”

• cites [65] Evans, 2013, Networks of anatomical covariance, Neuroimage sourceNot checked yet

  Not checked in the free run.