Using new tract-based analysis for DTI to assay structural connectivity in GAD and healthy controls, we observed bilaterally reduced FA of the uncinate fasciculus, a prominent white matter pathway connecting ventral PFC and ACC regions to the amygdala and other limbic areas. This effect was observed in the full sample of 49 GAD patients with and without current comorbid Axis I conditions and 39 healthy controls, and was particularly pronounced for patients with no comorbidities. These uncinate fasciculus findings suggest a structural basis for emotion regulation deficits in GAD6–9
and are consistent with previous functional imaging reports of abnormal activation patterns in the amygdala and ACC.4,7–13
Analyses conducted across imaging modalities elucidated the functional significance of the structural differences in the uncinate fasciculus. Across all subjects, lower FA values were associated with reduced negative coupling between the ACC and amygdala, precisely the relationship expected for poorer regulatory function. Decreased structural integrity of the uncinate fasciculus in patients with GAD may have detrimental functional consequences for emotion regulation, thereby contributing to heightened anxiety.
These DTI findings are the first report of uncinate fasciculus abnormalities in GAD. Of note, a recent DTI study of GAD did not report on FA for the uncinate fasciculus, but instead used a method assessing apparent diffusion coefficient (ADC), which is equivalent to MD, for circular ROIs in each of the four major brain lobes and the corpus callosum, and found no group differences for either the frontal or temporal lobe.42
The uncinate fasciculus findings here provide complementary support for past fMRI studies in GAD that have noted hyperactivity of the amygdala relative to healthy controls while participants are involved in processes such as implicit emotion regulation and conflict monitoring,9
the anticipation of emotional (and non-emotional) images,4
and viewing emotional faces.10,11
One interpretation of the amygdala hyperactivity that has frequently been observed in GAD is that patients fail to effectively recruit prefrontal circuitry that serves to regulate amygdala responses. Strong support for this hypothesis comes from the work of Etkin and colleagues,9
who demonstrated decreased coupling of the pregenual ACC and amygdala during the implicit regulation of emotional conflict in GAD. Such decreased coupling may be due to reductions in the integrity of the uncinate fasciculus, which is the primary white matter pathway connecting ventral prefrontal cortex with limbic structures, including the amygdala.31,36
By relating DTI data for the uncinate fasciculus to fMRI data on a disorder-relevant task of anticipatory function, findings here provide evidence that reduced microstructural integrity of this pathway is likely to have functional consequences for prefrontal-limbic communication. This builds on two important earlier reports that investigated relations between DTI and fMRI data related to anxiety and mood disorders.39,40
Our analytic procedure reduced a multi-step procedure for examining relations among DTI, fMRI, and psychopathology criterion (bipolar disorder39
, trait anxiety40
) to a single step that incorporates all three domains and allows for the simultaneous assessment of uncinate fasciculus structural integrity and diagnostic group (and their interaction) in predicting context-dependent functional connectivity with the amygdala.
Our data suggest that uncinate fasciculus integrity may be central to the previous observation in GAD of reduced functional connectivity between the pregenual ACC and amygdala.9
Indeed, in an analysis analogous to that conducted by Etkin et al.9
with group as the sole predictor of context-dependent connectivity, GAD patients showed reduced connectivity between the pregenual ACC and the amygdala (Author Supplemental Figure 2
). This group main effect was not significant in our primary model that included both group and uncinate fasciculus values as predictors, reflecting the substantial overlap between those individuals with GAD and those with the lowest FA values. Findings from both of these studies demonstrate decreased connectivity between the pregenual ACC and amygdala in GAD, with the current report emphasizing the importance of structural contributions. In addition, the current report extends previous findings9
of altered functional connectivity to the domain of anticipatory processing. This replication across experimental paradigms provides evidence that altered pregenual ACC – amygdala circuitry may be central to the pathology of GAD.
These structural and functional imaging studies point to a neurobiological basis for deficient emotion regulation abilities in individuals with GAD. Studies examining voluntary emotion regulation frequently report activation in many regions of the PFC and ACC, which is often inversely related to amygdala activation.25,26,28,29,43,80,81
During the anticipation of aversive images, we identified negative functional coupling of the pregenual ACC and amygdala only
in those subjects with the highest uncinate fasciculus values. Despite the lack of explicit task instructions, it seems likely that participants nevertheless enacted preparatory regulatory strategies during the anticipation period. Our data suggest that decreased uncinate fasciculus integrity in GAD may interfere with this prefrontal regulation of amygdala activation, adding to a growing literature on altered prefrontal-amygdala communication in GAD.4,9,12
In addition to these pregenual ACC findings, subjects with higher uncinate fasciculus values also showed greater negative coupling between the dorsolateral PFC and amygdala during the anticipation of aversion. Of note, there are robust connections between the amygdala and ventral portions of the PFC/ACC,36,82,83
while more dorsal portions of the PFC project weakly or not at all to the amygdala.36,84,85
It may be that ventral portions of the PFC/ACC serve as critical nodes in facilitating communication between dorsal PFC regions and the amygdala during regulation of emotional responses.29,30,86
Future research could test the hypothesis that deficient performance in GAD on an explicit emotion regulation task previously shown to engage the dorsolateral PFC25,26
is mediated by reduced integrity of the uncinate fasciculus.
Specificity of the findings was addressed in three ways. First, the group differences were anatomically specific to the uncinate fasciculus, as indicated by the absence of group differences elsewhere in the brain. This was determined using tract-based analyses in the cingulum, corpus callosum and inferior fronto-occipital fasciculus, as well as voxelwise, whole-brain analyses.
Second, the group differences for the uncinate fasciculus were strongest for the GAD patients without comorbidities, suggesting some degree of specificity for GAD. This observation stands in contrast to the identification of reduced uncinate fasciculus FA in social anxiety disorder,38
and bipolar disorder.39
Indeed, the accumulating positive findings across different studies suggest that decreased integrity of the uncinate fasciculus may be a general risk factor for affective pathology. Future research investigating questions of comorbidity and specificity might focus in particular on uncinate fasciculus structure in unipolar depression, as 23 of 28 subjects in our comorbid group had a current or past diagnosis of MDD.
Third, group differences were observed for FA but not MD. The null findings for MD in the uncinate fasciculus are consistent with the only previously published MD findings for GAD.42
Although the precise biological characteristics associated with different DTI measures are not fully known,63
FA and MD likely quantify complementary aspects of brain microstructure. Differences in FA may reflect alterations in either myelination or axonal density, whereas MD reflects the overall density of tissue membranes irrespective of fiber orientation.63
Accordingly, findings here for FA implicate a difference in the microstructural components that have directional dependence due to myelination or axonal density. Of note, the evidence for minimal axonal plasticity in the adult brain87,88
is relevant to findings for the present sample, which included a broad age range. The role of uncinate fasciculus structure in the development and course of GAD and other affective disorders are important topics for future investigations.
Of potential relevance to the etiology of GAD, ancillary analyses examined relations between uncinate fasciculus integrity and common genetic polymorphisms linked to anxiety. We did not replicate recent findings of reduced uncinate fasciculus FA in healthy volunteers for the low-expressing 5-HTTLPR allele,41
although this pattern was observed for the patients. We also failed to replicate the finding of reduced uncinate fasciculus FA for the BDNF Met allele.53
Further research is needed to determine the replicability of that original finding for the BDNF Met allele53
and to clarify whether the effects of this polymorphism on anxiety52
and fear extinction51,53
are mediated by the uncinate fasciculus or a separate mechanism. A critical consideration is that while the sample of 88 subjects is large for a neuroimaging patient study, and larger than for many published neuroimaging genetics studies, the sample size is insufficient for detecting the smaller effect sizes that are typical of genetics studies; thus, these mainly negative genetics findings are not conclusive.50
In summary, using DTI tract-based analysis, we identified evidence of reduced integrity of the uncinate fasciculus, a crucial white matter pathway linking ventral PFC and ACC to limbic regions, in GAD. These results indicate that altered structure of a neural pathway involved both in normative emotion regulation and fear extinction processes may contribute to atypical emotional processing in GAD. The group differences in uncinate fasciculus structural connectivity, together with the observed association with functional connectivity, support a model positing emotion regulation deficits in GAD6–8
and suggest weak top-down control of amygdala reactivity. Further research is needed to determine how worry, the hallmark feature of GAD, affect the neurobiology identified here, but its presumed function in avoiding negative emotional experiences may actually sensitize amygdala activity, resulting in a generalized state of heightened anxiety.89
Finally, the identification of a relationship between measures of structural and functional connectivity in a circuit highly relevant for emotion regulation and anxiety disorders underscores the potential and promise for new discovery that can come about through the integration of independent modalities of imaging data.