Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 1.
Published in final edited form as: Autism Res. 2021 Jul 15;14(10):2100–2112. doi: 10.1002/aur.2579

Functional connectivity within an anxiety network and associations with anxiety symptom severity in middle-aged adults with and without autism

R Tung 1,2,*,#, M A Reiter 1,2,3,*,#, A Linke 1,2, JS Kohli 1,2,3, M K Kinnear 1,2, R-A Müller 1,2,3, R A Carper 1,2,3,+
PMCID: PMC8487970  NIHMSID: NIHMS1722530  PMID: 34264028

Abstract

Anxiety is highly prevalent in autism spectrum disorders (ASDs). However, few functional MRI (fMRI) studies of ASDs have focused on anxiety (and fewer still on anxiety in middle-aged adults). Thus, relationships between atypical connectivity and anxiety in this population are poorly understood. The current study contrasted functional connectivity within anxiety network regions across adults (40-64 years) with and without autism, and tested for group by functional connectivity interactions on anxiety. 22 adults with ASDs (16 males) and 26 Typical Control (TC) adults (22 males) completed the Beck Anxiety Inventory and a resting-state fMRI scan. An anxiety network consisting of 12 regions of interest was defined, based on a meta-analysis in TC individuals and 2 studies on anxiety in ASDs. We tested for main effects of group and group by anxiety interactions on connectivity within this anxiety network, controlling for head motion using ANCOVA. Results are reported at an FDR adjusted threshold of q < .1 (corrected) and p < .05 (uncorrected). Adults with ASDs showed higher anxiety and underconnectivity within the anxiety network, mostly involving bilateral insula. Connectivity within the anxiety network in the ASD group showed distinct relationships with anxiety symptoms that did not relate to ASD symptom severity. Functional connectivity involving the bilateral posterior insula was positively correlated with anxiety in the ASD (but not the TC) group. Increased anxiety in middle-aged adults with ASD is associated with atypical functional connectivity, predominantly involving bilateral insula. Results were not related to ASD symptom severity suggesting independence of anxiety-related effects.

Keywords: Anxiety, autism, ASD, adults, functional connectivity, resting state fMRI

Lay Summary

Anxiety is very common in adults with autism but the brain basis of this difference is not well understood. We compared functional connectivity between anxiety-related brain regions in middle-aged adults with and without autism. Adults with autism were more anxious and showed weaker functional connections between these regions. Some relationships between functional connectivity and higher anxiety were specific to the autism group. Results suggest that anxiety functions differently in autism.

1. Introduction

Anxiety disorders are frequently observed in adults with autism spectrum disorders (ASDs), with lifetime and current prevalence estimates at over 40%, and 25%, respectively (Hollocks, Lerh, Magiati, Meiser-Stedman, & Brugha, 2019). Indeed, adults with ASDs have been shown to exhibit higher rates of nearly all anxiety-related disorders compared to typical control (TC) adults (Nimmo-Smith et al., 2020) or adults with other developmental disabilities (Gillott & Standen, 2007). In ASDs, anxiety isn’t purely a function of symptom severity, with susceptibility highest among adults without co-occurring intellectual disability (rather than those with the most severe ASD symptoms) as seen in a large population-based study (Nimmo-Smith et al., 2020), although a meta-analysis based on smaller clinical samples had mixed findings (Hollocks et al., 2019). Anxiety is associated with poor emotional regulation in children, reduced life satisfaction, greater social difficulties during the transition into early adulthood, and significantly lowered quality of life in ASD (Gotham, Brunwasser, & Lord, 2015; van Heijst & Geurts, 2015; van Steensel, Bögels, & Dirksen, 2012).

During middle-age and aging in adults without ASDs, anxiety disorders remain among the more common psychological disorders, although some anxiety symptoms may diminish (Byers, Yaffe, Covinsky, Friedman, & Bruce, 2010; Lenze & Wetherell, 2011). However, it is unknown whether anxiety symptoms change with aging in ASDs (Davis et al., 2011; Lever & Geurts, 2016; Uljarević et al., 2020). Differences in neurocircuitry may underlie the elevated anxiety rates in ASDs, and network reorganization that occurs during healthy aging (Geerligs, Renken, Saliasi, Maurits, & Lorist, 2015; He et al., 2014; Onoda, Ishihara, & Yamaguchi, 2012; Tomasi & Volkow, 2012; Varangis, Habeck, Razlighi, & Stern, 2019) may play out differently against this foundation. During typical aging, studies have shown declining intra-network connectivity and network segregation, affecting default mode (Geerligs et al., 2015; Onoda et al., 2012; Tomasi & Volkow, 2012), salience (He et al., 2014; Onoda et al., 2012), fronto-parietal (Geerligs et al., 2015), cingulo-opercular (Geerligs et al., 2015; Varangis et al., 2019), and dorsal attention networks (Tomasi & Volkow, 2012; Varangis et al., 2019). Such changes in neurotypical aging may be compounded by pre-existing atypical network organization in ASDs earlier in life and potentially alter rates of cognitive or affective change. Thus, we cannot assume that autism research on the neural substrates of anxiety conducted in youth and early adulthood generalizes to middle age and aging. Investigation at the level of distributed neural networks and their connectivity is required within this age-range, and such research may eventually help us understand how increased vulnerability to anxiety in ASDs could be related to neurobiological functioning of anxiety circuits.

Neural substrates of anxiety

Resting-state fMRI has become a method of choice to analyze intrinsic functional connectivity in the absence of a specific task, and has been widely used to study anxiety in TC populations. While large-scale meta-analyses defining anxiety circuits exist for TC populations (Etkin & Wager, 2007; Weber-Goericke & Muehlhan, 2019), research on anxiety circuits in ASDs is far more limited leaving uncertainty as to the full extent of the regions affected. One reasonable approach to this uncertainty is to define a ‘broad anxiety network’ incorporating regions highlighted by studies of anxiety and anxiety disorders without co-occurring ASD, (particularly anxiety disorders most prevalent in ASDs), and to also include any further regions that have emerged in the more limited literature on anxiety in ASDs. This approach was chosen in the current study.

Social anxiety is the most prevalent anxiety disorder diagnosed in ASDs according to a meta-analysis by Hollocks et al. (2019). In TC adults, implicated neural regions include the amygdala, parahippocampal gyrus, fusiform gyrus, globus pallidus, anterior and posterior insula, inferior frontal gyrus, and superior temporal gyrus (see meta-analysis by Etkin & Wager, 2007). Notably, many of these regions also play roles in other anxiety disorders, including generalized anxiety disorder [amygdala, anterior insula, prefrontal cortex (Mochcovitch, da Rocha Freire, Garcia, & Nardi, 2014)], panic disorder [superior temporal gyrus, hippocampus, amygdala, prefrontal cortex (de Carvalho et al., 2010)], obsessive compulsive disorder [amygdala, parahippocampal gyrus, insula, inferior frontal cortex (Nakao, Okada, & Kanba, 2014)], and specific phobias [amygdala, fusiform gyrus, insula (Etkin & Wager, 2007)]. Different regions of the brain may be particularly informative to certain anxiety disorders (e.g. hyperactivation of the amygdala and insula are more commonly observed in social anxiety disorder and specific phobia than in Post-Traumatic Stress Disorder (Etkin & Wager, 2007) and similar topological variation may be found for anxiety co-occurring with ASDs. Thus, improved understanding of anxiety in ASDs requires investigation of interactions between a broader array of regions (including interactions between multiple regions beyond the amygdala as potentially part of a ‘broad anxiety network’). To our knowledge no study to date has attempted to study functional connectivity within a broad anxiety network in adults with ASDs.

Neural substrates of anxiety in ASDs

What little is known about the neural substrates of anxiety in ASDs is derived almost entirely from studies in children and adolescents; no studies to date have focused on adults past the fourth decade of life, although most individuals with ASDs continue to meet diagnostic criteria for autism in adulthood (Magiati, Tay, & Howlin, 2014). One study of adolescents with ASDs and co-occurring anxiety tested the whole-brain response during reward anticipation and negative feedback. Findings from a comparison to adolescents with ASD who had no co-occurring anxiety indicated involvement of insular and frontal cortices with higher right insula activity when anticipating reward, and lower medial and lateral prefrontal cortex activation when receiving negative feedback (Mikita et al., 2016). This suggests that insular, and medial and lateral prefrontal cortices may have atypical involvement in the anxiety network in ASDs.

Numerous studies have found the amygdala to be affected in anxiety without co-occurring ASD, but also in ASD independently of anxiety (Amaral & Corbett, 2003), highlighting its probable and crucial role in transdiagnostic processes underlying co-occurring conditions (Hennessey, Andari, & Rainnie, 2018). In children with ASDs, anxiety has been related to atypical volume (Herrington et al., 2017) and activation (Juranek et al., 2006) of the amygdala. Increased amygdala activation and reduced frontal activation have also been observed in response to emotional facial expressions, an effect associated with greater social anxiety (Kleinhans et al., 2010). Atypical functional connectivity of amygdala-frontal circuits has further been associated with anxiety in TC populations [e.g. see systematic review and meta-analysis by (Kolesar, Bilevicius, Wilson, & Kornelsen, 2019)]. These results suggest some similarity in mechanisms implicated in anxiety across ASD and TC populations. However, other studies suggest distinct neural mechanisms related to anxiety in ASDs involving the amygdala. For example, amygdala connectivity has been shown to be differentially associated with anxiety in ASDs compared to TC adolescents (Kleinhans et al., 2016). Additionally, young adults with ASDs have been found to show atypical amygdala response to fear conditioning paradigms (Top Jr. et al., 2016). Thus, although some similar regions and circuits appear to be involved in anxiety in ASD and TC populations, exact substrates may very well diverge.

The neural substrates underlying co-occurring anxiety and ASDs have yet to be fully examined. At the same time, findings on the severity of anxiety symptoms in middle-age and aging in ASD are uncertain. To address this gap in the literature the main goals of this study were (i) to contrast adults with and without ASDs in regard to functional connectivity within a broad network of known anxiety related regions, (ii) to determine whether relationships between functional connectivity and anxiety symptom severity differ between groups, and (iii) to test whether atypical connectivity in these circuits was related to core autism symptoms rather than to anxiety.

2. Methods

2.1. Participants

29 adults with ASDs and 38 typical control (TC) participants aged 40 to 65 years were recruited through community advertising as part of an ongoing longitudinal study. Adults with ASD-related medical conditions (e.g. Fragile-X syndrome, tuberous sclerosis, epilepsy) or other neurological conditions (e.g. Tourette syndrome) were excluded. Exclusionary criteria for the TC group included personal or family history (first degree relatives) of ASDs, neurological disorders, schizophrenia, bipolar disorder, severe major depressive disorder, obsessive-compulsive disorder, psychosis, or other developmental disorders. Participants were not excluded from the ASD or TC group for taking psychotropic medications (including medications for anxiety, depression, and other medical conditions) in order to maintain ecological validity, as research has shown that over 50% of adults with ASDs take at least one psychotropic medication (Buck et al., 2014; Cvejic, Arnold, Foley, & Trollor, 2018). Supplementary Table S1 depicts medication use within the current study’s sample. Participants were not asked to abstain from medication use during the MRI portion of the study for ethical reasons, and behavioral and psychodiagnostic assessments were also completed while participants were taking their medications as usual. Three ASD participants were excluded for unusable MRI data and one for not fully meeting diagnostic criteria. One TC participant was excluded from analysis due to incidental finding of brain dysmorphology and two for unusable MRI data. Group matching was finalized prior to any analyses of the neuroimaging data (except quantification of head motion) to prevent selection bias. In order to match groups on head motion and age, individual TC participants were excluded (one at a time) and descriptive statistics (mean, standard deviation, and range) on the matching variables were reviewed manually following each exclusion. A total of 9 TC participants were excluded. Exclusion of an additional 3 ASD participants was required to suitably match groups on age and head motion. The final sample included 22 ASD and 26 TC participants. Demographics for all participants included in analyses are summarized in Table 1.

Table 1.

Demographic information for all participants

ASD (n = 22) TC (n = 26)
Mean ± standard deviation (range) Mean ± standard deviation (range)
Age (years) 49.45±6.03 (40-60) 51.00±7.03 (40-64) t(46)=.81, p=.422
Sex 6 female 4 female --
Handedness 2 left-handed 3 left-handed --
Total brain volume (cm3) 1115.57±90.19 (875.67-1244.79) 1160.975±90.46 (992.20-1321.40) t(43) =−1.68, p=.10
Full Scale IQ 104.32±23.41 (58-143) 116.4±11.01 (90-138) t(28.75)=2.23, p=.03
Verbal Comprehension 102.86±25.88 (45 - 160) 116.3±13.13 (96-144) t(29.96)=2.21, p=.035
Perceptual Reasoning 105.09± 22.40 (63 - 138) 112.62±13.44 (75-138) t(32.92)=1.38, p=.176
ADOS-2 Social Affect + Restricted Repetitive Behavior Total 13.90±4.37 (7-23) -- --
ADOS-2 Social Affect Total 10.76±4.11 (5-19) -- --
ADOS-2 Restricted Repetitive Behavior Total 3.14± 1.85 (1-8) -- --
BAI 9.84±7.74 (0-27) 1.85±2.74 (0-12) t(21.31)=−4.31, p=.0003
Head motion Run 1 (RMSD) .104±.044 (.05-.183) .087±.033 (.048-.187) t(46)=−1.536, p=.131
Head motion Run 2
(RMSD)		 .106±.042 (.053-.211) .091±.040 (.048 - .199) t(46)=−1.270, p=.210
Mean HR (BPM) 66.50±11.15 (38.01-88.54) 64.85±11.12 (46.26-83.98) t(45)=−.51, p=.61
SD HR (BPM) 2.21±1.08 (.6-4.55) 1.88±.71 (.77-3.73) t(45)=−1.27, p=.21
Mean RVT .45±0.19 (.14-.81) .54±.17, (.20-.75) t(45)=1.71, p=.09
SD RVT .14±.05 (.07-.25) .12±.036 (.08-.22) t(45)=−1.24, p=.21
Open in a new tab

RMSD=root mean square difference; HR=heart rate, RVT=respiratory volume per time, mean SD (range) reported for age, total brain volume, RMSD, HR and RVT, TBV unavailable for n=2 TC and n=1 ASD participants, HR unavailable for n=1 TC participants, RVT unavailable for n=1 TC participant

Informed consent was acquired from participants or their conservators, and all participants were compensated for their time. The study protocol was approved by the institutional review boards of the University of California San Diego and San Diego State University.

2.2. Diagnostic and Neuropsychological Assessments

ASD diagnoses were confirmed by a clinical psychologist in accordance with Diagnostic and Statistical Manual of Mental Disorders 5th edition (DSM-5) criteria (American Psychiatric Association, 2013) and supported by module 4 of the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) (Lord et al., 2014), and developmental history from review of records and/or developmental interview with the participant or a family member. These latter interviews and reviews focused on establishing the presence of autism symptoms during childhood (e.g. language delay, social deficits). IQ was assessed using the Wechsler Abbreviated Scale of Intelligence, 2nd edition (WASI-II) (Wechsler, 2011). Anxiety levels were measured using the Beck Anxiety Inventory (BAI) (Steer & Beck, 1997).

2.3. Magnetic Resonance Imaging

MRI data were collected at the University of California San Diego Center for Functional MRI on a 3T GE MR750 using a Nova Medical 32 channel head coil. A fast 3D spoiled gradient recalled (FSPGR) T1-weighted sequence was used to acquire high-resolution structural images (0.8 mm isotropic voxel size, NEX=1, TE/TI=min full/1060ms, flip angle 8°, FOV=25.6cm, matrix=320×320, receiver bandwidth 31.25hz). Motion during structural acquisitions was corrected in real-time using three navigator scans (PROMO, real-time prospective motion correction; (White et al., 2010), and images were bias corrected using the GE PURE option. A multiband EPI sequence which allows simultaneous acquisition of multiple slices was used to acquire two fMRI runs (6 minute duration each) with high spatial and temporal resolution (TR=800ms, TE=35ms, flip angle 52°, 72 slices, multiband acceleration factor 8, 2 mm isotropic voxel size, 104×104 matrix size, FOV 20.8cm, 400 volumes per run). Two separate 20s spin-echo EPI sequences with opposing phase encoding directions were also acquired using the same matrix size, FOV and prescription to correct for susceptibility-induced distortions. Before the functional scans, participants received the instruction: “Keep your eyes on the cross. Let your mind wander, relax, but please stay as still as you can. Try not to fall asleep.” Adherence to the instructions to remain awake with eyes open was monitored with an MR-compatible video camera. Scans were repeated or aborted if participants did not follow the instructions to fixate. Two participants (both in the TC group) had to be excluded from analyses because they could not keep their eyes open during the resting state scans even after repeat attempts and reminders. Heart rate and respiration during the functional scans were recorded continuously using a Biopac pulse oximeter and respiratory belt. While we attempted to collect this data from all participants for the full imaging period, not all participants had high quality heart rate and respiratory recordings for the full duration of the two resting state fMRI runs (e.g., due to signal loss for the pulse oximetry and/or respiratory belt recordings. Available physiological recordings during the fMRI scans were analyzed with the PhysIO toolbox (Kasper et al., 2017). The mean and standard deviation of the heart rate and respiratory volume per time were calculated for each participant and compared across groups using independent samples t-test (see Table 1). Since groups did not differ on any of physiological measures, they were not included in further analyses.

2.4. Data preprocessing and denoising

MRI data were preprocessed, denoised and analyzed in Matlab 2016b (Mathworks Inc., Natick, MA, USA) using SPM12 (Wellcome Trust Centre for Neuroimaging, University College London, UK) and the CONN toolbox v17f. Structural images were converted from DICOM to NIfTI format and coregistered to the mean functional image, segmented and normalized to MNI space using non-linear registration and the default tissue probability maps included in SPM12. White matter (WM) probability maps obtained from segmentation of the structural image for each individual subject were thresholded at .99 and eroded by 1 voxel. Due to high variability in ventricle size across participants, the template CSF map was used to extract CSF time courses for all participants (thresholded at .5 and eroded by 1 voxel). There was no overlap with gray matter voxels for the eroded WM and CSF masks for any participant. WM and CSF time courses were extracted from the thresholded and eroded masks using aCompCor (Behzadi, Restom, Liau, & Liu, 2007) for subsequent nuisance regression (see below).

Functional images were corrected for susceptibility-induced distortions using the two spin-echo EPI acquisitions with opposite phase encoding directions and FSL’s TOPUP tools (Smith et al., 2004). Subsequently, functional images were motion-corrected using rigid-body realignment as implemented in SPM12. The Artifact Detection Toolbox (ART, as installed with conn v17f) was used to identify outliers in the functional image time series from the resulting 6 motion parameters (3 translational and 3 rotational) that had frame-wise displacement (FD) >0.9mm and/or changes in signal intensity that were greater than five standard deviations. As oscillations due to respiration are prominent in motion parameters derived from multiband EPI realignment (Fair et al., 2020) and would result in unnecessary censoring of large chunks of data in some participants, the thresholds to detect outliers were more lenient than those used for standard resting state fMRI acquisitions with slower TRs. In order to ensure that none of our findings were due to differences in apparent motion between groups, participants were matched on RMSD calculated from rigid-body realignment of the raw data prior to TOPUP correction (Table 1). ASD and TC group did not differ in the number of time points detected as outliers (run1: t(46)=−1.8, p=.08; run2: t(47)=−1.53, p=.13), with a maximum of 5% of volumes detected as outliers in the individual runs (range ASD, run1: 0-3.5%, run2: 0-4%; range TC, run1: 0-2.25%, run2: 0-5%)

Functional images were directly normalized to MNI space with the same non-linear registration as used for the structural images. Since all analyses were run on averaged voxel time series within pre-defined ROIs, no prior smoothing was applied to the data. Voxel time series were normalized to percent-signal change in each run. Band-pass filtering using a temporal filter of 0.008 to 0.08 Hz was carried out as part of the nuisance regression (“simult” option in the conn toolbox) which also included scrubbing of the motion outliers detected by the ART toolbox, and regression of the 6 motion parameters and their derivatives, as well as the first five PCA component time series derived from the CSF and white matter masks. The residuals of the nuisance regression were then used for all subsequent functional connectivity analyses with timeseries concatenated across the two runs.

2.5. Functional Connectivity Analysis

2.5.1. Regions of interest

For the functional connectivity analyses, regions of interest (ROIs) were created in the CONN toolbox using MNI coordinates of seven cortical and three subcortical ROIs implicated in a meta-analysis of task-based functional activity studies on anxiety in individuals without ASDs (Etkin & Wagner, 2007). These ROIs were located in the right fusiform gyrus, insulae (left anterior and bilateral posterior ROIs), right inferior frontal gyrus, right parahippocampal gyrus, left superior temporal gyrus, bilateral amygdalae, and left globus pallidus. Additionally, ventro-medial prefrontal cortex (vmPFC) and right middle frontal cortex (MFC) ROIs were included as these regions have received specific attention in the literature on neural substrates of anxiety in ASDs (Kleinhans et al., 2016; Mikita et al., 2016). Cortical ROIs were 10mm diameter spheres, while 8mm diameter spheres were used for the subcortical regions to avoid capturing signal from adjacent regions. See Figure 1 and Supplementary Table S2 for ROI locations. The ‘anxiety network’ henceforth discussed in this manuscript refers to the collection of ROIs included in the current study, a network of regions shown to be involved in the anxiety response across a set of different anxiety disorders in non-ASD populations. We grouped these regions together based on the previous literature, and reiterate that this anxiety network was not defined using data-driven approaches such as ICA.

Figure 1: Glass brain representation of the 12 ROIs.

Figure 1:

Open in a new tab

1. L.A = Left Amygdala, 2. R.A = Right Amygdala, 3. L.GP = Left Globus Pallidus, 4. Ant-L.I = Anterior Left Insula, 5. Post-L.I = Posterior Left Insula, 6. Post-R.I = Posterior Right Insula, 7. R.IFG = Right Inferior Frontal Gyrus, 8. R.FG = Right Fusiform Gyrus, 9. R.PHG = Right Parahippocampal Gyrus, 10. L.STG = Superior Temporal Gyrus, 11. L.vmPFC = Left Ventro-medial Prefrontal Cortex, 12. R.MFC = Right Middle Frontal Cortex

2.5.2. Time series analysis

BOLD time series were averaged across all voxels within each ROI, and functional connectivity between all 66 ROI-ROI pairs was estimated using bivariate Pearson correlation standardized with a Fisher z-transform. Each of these was entered into a general linear model with group, anxiety (BAI total score), and group-by-anxiety interaction as predictor variables and controlling for in-scanner head motion (RMSD) as a nuisance variable. Thus, a total of 132 coefficients were calculated in the current study (66 main effects, 66 interaction effects). False Discovery Rate correction was used to adjust all 132 resulting p-values (Storey, 2002). Given sample size limitations, the significance threshold was set at q<.1 (FDR corrected), p<.05 (uncorrected).

3. Results

3.1. Anxiety levels in adults with and without ASDs

Anxiety levels were indexed by the BAI Total score with 40.9% of participants with ASDs scoring in the ‘Mild to Moderate’ or ‘Moderate to Severe’ range, while only a single participant in the TC group measured outside of the ‘Minimal Anxiety’ range. (Research has shown that, on average, healthy adults obtain BAI scores in the ‘Minimal Anxiety’ range, with an average BAI score of 6 and a median BAI score of 4 (Crawford, Cayley, Lovibond, Wilson, & Hartley, 2011)). Overall, the ASD group showed significantly higher anxiety scores (BAI Total score) compared to the TC group (t(43)=4.88, p<.001; Figure 2). There were no significant correlations between anxiety levels and age in either the ASD group (r=−.25, p=.32) or the TC group (r=−.06, p=.77), nor between anxiety levels and autism symptom severity as measured by ADOS-2 Total scores in the ASD group (r=−.26, p=.22).

Figure 2: BAI score distribution for ASD and TC samples.

Figure 2:

Open in a new tab

The ASD group (left) shows significantly higher anxiety levels compared to the TC group (right), t(43)=4.88, p<.001. Dotted orange lines denote clinical cut off levels for levels of anxiety on the Beck Anxiety Inventory (BAI).

3.2. Functional connectivity of the anxiety network

We found significantly reduced functional connectivity within the anxiety network in the ASD compared to the TC participants involving four ROI-ROI pairings (lower positive correlations between BOLD signals in two ROIs, see Table 2). The ROI pairs were the posterior left insula-right inferior frontal gyrus (Hedge’s g = 0.29), posterior left insula-right fusiform gyrus (Hedge’s g = 0.63), posterior right insula-right fusiform gyrus (Hedge’s g = 0.35), and posterior right insula-left superior temporal gyrus (Hedge’s g = 0.35). These differences were independent of head-motion. In addition, for eight ROI-ROI pairings the relationship between anxiety symptom severity (BAI scores) and functional connectivity was moderated by group (interaction effects). These pairs were the left amygdala-right parahippocampal gyrus, right amygdala-right fusiform gyrus, left globus pallidus-right middle frontal cortex, posterior left insula-posterior right insula, posterior left insula-right inferior frontal gyrus, posterior right insula-right inferior frontal gyrus, posterior right insula-right fusiform gyrus, and left ventro-medial prefrontal gyrus-right middle frontal cortex. However, caution is warranted given the lower variability in BAI scores in the TC group. Table 2 lists the statistical values for ROI-ROI pairings with main effects and/or interactions that survive FDR correction. See Supplementary Table S3 showing results that did not reach statistical significance and for uncorrected p-values. Statistically significant results are visualized in Figure 3; the remainder of the ROI-ROI results are visualized in supplementary materials (Supplementary Figure S1).

Table 2.

Effects of group and anxiety on functional connectivity

ROI – ROI Pair Mean FC ASD (Standard Deviation) Mean FC TC (Standard Deviation) ME β ME Standard Error ME p value (FDR corrected p) Group*Anxiety Interaction β Group*Anxiety Interaction Standard Error Group*Anxiety Interaction p value (FDR corrected p value)
L.A – R.PHG .38 (.15) .38 (.19) −.09 .07 .20 (.38) .03 .01 .02* (.09)**
R.A – R.FG .18 (.17) .12 (.24) −.001 .09 .99 (.70) .04 .02 .01* (.09)**
L.GP – R.MFC .12 (.14) .11 (.14) .02 .05 .77 (.66) −.02 .01 .03* (.0997)**
Post.L.Ins – Post.R.Ins .69 (.34) .73 (.26) −.33 .12 .01* (.1) .06 .02 .01* (.08)**
Post.L.Ins – R.IFG .42 (.27) .49 (.21) −.31 .09 .002* (.08)** .05 .02 .004* (.095)**
Post.L.Ins – R.FG .26 (.19) .37 (.16) −.26 .07 .001* (.08)** .03 .01 .03* (.102)
Post.R.Ins – R.IFG .33 (.21) .36 (.24) −.22 .09 .02* (.11) .05 .02 .01* (.0997)**
Post.R.Ins – R.FG .28 (.17) .35 (.22) −.23 .09 .01* (.08)** .04 .02 .01* (.08)**
Post.R.Ins – L.STG .29 (.26) .38 (.26) −.30 .11 .01* (.09)** .03 .02 .17 (.36)
L.vmPFC-R.MFC .23 (.21) .22 (.24) −.16 .10 .11 (.30) .04 .02 .01* (.08)**
Open in a new tab

Only effects that survive FDR correction are shown

*

Uncorrected p-values significant at p < .05

**

FDR corrected p-values significant at q < .1

FDR corrected p-values bolded

ME = Main Effect of Group

FC = Functional Connectivity

Figure 3: Group differences in anxiety network functional connectivity, group by anxiety severity interactions on functional connectivity within the anxiety network, and correlations between anxiety and functional connectivity within the anxiety network in ASDs.

Figure 3:

Open in a new tab

Top: A. Functional connectivity (z-transformed correlations) for ROI pairs in TC (left) and ASD (right). Overlay triangles: Cyan fill = significant effect of group on connectivity (main effect); Bold outline = significant group by anxiety interaction on connectivity. B. Glass brain visualizes significant functional connectivity differences between groups (blue lines represent ASD < TC).

Bottom: C. Correlation matrices comparing main effect of group, group x BAI interaction and BAI/functional connectivity correlations. Plus signs (+) represent significant interactions, triangles (▼) represent significant group differences and asterisks (*) represent significant BAI/functional connectivity correlations

All highlighted results FDR corrected at q<.1.

3.3. Post-hoc analysis: convergent and divergent validity

In order to better interpret the implications of observed functional connectivity differences within the anxiety network in adults with ASDs, we conducted post-hoc correlational analyses after separating the two subject groups. These analyses aimed to characterize relationships between functional connectivity within the anxiety network and three variables potentially related to connectivity within this network [1] anxiety symptom severity (BAI total), [2] autism symptom severity (ADOS-2 total), and: [3] full-scale IQ. Eighteen ROI-ROI pairings showed positive correlations between strength of functional connectivity and anxiety symptom severity in the ASD group after FDR correction (Storey, 2002). These were the left globus pallidus-posterior right insula, left globus pallidus-right fusiform gyrus, left globus pallidus-right parahippocampal gyrus, left globus pallidus-superior temporal gyrus, anterior left insula-posterior left insula, anterior left insula-posterior right insula, anterior left insula-right fusiform gyrus, anterior left insula-right parahippocampal gyrus, anterior left insula-superior temporal gyrus, posterior left insula-posterior right insula, posterior left insula-right inferior frontal gyrus, posterior left insula-right fusiform gyrus, posterior left insula-right parahippocampal gyrus, posterior left insula-superior temporal gyrus, posterior right insula-right inferior frontal gyrus, posterior right insula-right fusiform gyrus, posterior right insula-right middle frontal, right fusiform gyrus-superior temporal gyrus (Figure 3). One ROI-ROI pairing correlated with ADOS-2, one correlated with FSIQ in the ASD group, posterior right insula-left ventro-medial prefrontal cortex. In the TC group, only four correlated with FSIQ and only 2 ROI-ROI pairs survived multiple comparisons, anterior left insula-right middle frontal and inferior frontal gyrus to right middle frontal (Supplementary Table S4).

To examine possible differential effects in males and females, female participants (6 ASD, 4 TC) were excluded for additional analyses. Subsequently, only the main effect for posterior left insula-right inferior frontal gyrus and the group-by-anxiety interaction for posterior right insula-right inferior frontal gyrus remained significant, while an additional main effect for posterior left insula-left superior temporal gyrus (ASD < TC) was revealed (Supplementary Table S5). While several previous effects no longer reached statistical significance, likely due to the loss of statistical power, the directions of effects were unchanged such that the overall pattern of findings remained the same. We also tested for effects after removing 5 left-handed individuals, these results are available in Supplementary Table S6.

Lastly, the ASD and TC groups did not significantly differ in TBV (Table 1) but in order to rule out the possibility that varying degrees of atrophy in the studied age range might have confounded functional connectivity estimates, we: 1) compared the standard deviation (reflecting BOLD signal amplitude, e.g. (Bijsterbosch et al., 2017), each ROI between the two groups. Since BOLD signal amplitude is substantially higher in gray than white matter or CSF, any reduction in the gray matter volume fraction in a given ROI (e.g. due to atrophy) would be expected to decrease the standard deviation of the signal. Standard deviation of average time series within an ROI did not significantly differ between the ASD and TC groups for any ROI except the left globus pallidus, which had slightly higher standard deviation in the ASD (mean=0.30) than the TC (mean=0.26) group (p=.02, uncorrected for multiple comparison; all other p>.2). 2) Additionally, we assessed whether there were any correlations between TBV and functional connectivity estimates. Only one of 66 ROI pairs (left amygdala-right amygdala) showed a significant correlation (r=−.32, p=.03, uncorrected for multiple comparisons) with a mean r across ROI pairs of −.05 (mean |r|=.10). These results suggest that differences in atrophy in this age range did not systematically confound group differences in functional connectivity identified.

4. Discussion

This study examined intrinsic functional connectivity within an empirically defined anxiety network, comparing middle-aged adults with ASDs to a group of age, motion, gender, and FSIQ matched TC adults. Overall, we observed a pattern of reduced connectivity within the anxiety network in our sample of adults with ASDs, who also showed substantially higher anxiety levels compared to TC adults. Differences in functional connectivity most prominently involved insular ROIs bilaterally. Results of the follow-up correlational analyses within the ASD group provided reassurance that altered functional connectivity within the anxiety network was indeed related to anxiety rather than sociocommunicative and/or restrictive repetitive core symptomatology. Altered patterns of functional connectivity within the anxiety network that were observed in the current study may contribute to the higher prevalence of anxiety in ASDs, possibly reflecting a mechanism of biological vulnerability to anxiety in this population. However, causal relationships cannot be inferred due to the nature of the cross-sectional data collected, as atypical functional connectivity could have also developed as a result of chronic anxiety over the lifespan and further replication of these results in larger sample sizes is certainly warranted.

Brain-behavior links for anxiety in middle-aged adults with ASDs

A recent meta-analysis found that anxiety in TC children and adults is predominantly characterized by reduced within- and between-network intrinsic functional connectivity (Xu et al., 2019). Specific patterns associated with higher anxiety and involving the insula included decreased functional connectivity within the salience network, attenuated connectivity between salience and sensorimotor networks, and underconnectivity between superior temporal gyrus and anterior insula. The current study found reduced connectivity but increased anxiety in ASDs, based on which negative correlations between functional connectivity and anxiety could be predicted. Unexpectedly however, these correlations were positive in the ASD group. This paradoxical finding was also seemingly inconsistent with the negative associations between functional connectivity and anxiety observed in TC individuals by Xu et al. (2019), including similar ROI pairings (e.g., anterior insula and superior temporal gyrus).

Our results suggest that the relation between functional connectivity within the anxiety network and level of anxiety differs notably between adults with ASDs and TC adults. This is supported by several significant interaction effects, mostly involving bilateral insular connectivity and anxiety (Supplementary Figure S1). Some caution is needed in view of the limited range of anxiety scores in the TC group, which complicates interpretation of interaction effects. Nevertheless, the ASD-specific correlations between functional connectivity and anxiety observed here are notable, especially when viewed in the context of opposite brain-behavior relationships established in the broader TC anxiety literature. Moreover, in response to a reviewer’s suggestion, we tested for correlations between functional connectivity and anxiety across the entire ASD + TC sample (collapsing across groups). Of the 18 correlations between BAI and functional connectivity reported in the ASD group, none were significant after correction for multiple comparisons in the combined ASD + TC sample. These findings suggest not only atypically increased anxiety in middle-aged adults with ASDs, but also atypical anxiety-related brain-behavior links, which may help further define both the anxiety network itself as well as its mechanisms of operation in ASDs. Divergent functioning of the anxiety network may also lead to greater vulnerability to anxiety disorders in adults with ASDs, as previously suggested by Kleinhans et al. (2016). Differing neural substrates of anxiety may also have key treatment implications, especially for biological modes of intervention such as SSRIs, which are commonly used to treat anxiety (James et al., 2017).

Prominence of the insula in anxiety-related functional connectivity patterns in ASDs

The current study found main effects of group, interactions, and ASD-specific functional connectivity correlations with anxiety symptoms involving insula (predominantly posterior insula) and other anxiety-network ROIs. While the amygdala is perhaps the most frequently investigated region in human and non-human studies of anxiety, the insula has also received attention. The insula is functionally a highly diverse structure involved in somatosensory (including pain), chemosensory, auditory, interoceptive, autonomic, and social cognitive processes (Uddin, Nomi, Hébert-Seropian, Ghaziri, & Boucher, 2017). In addition, its role in anxiety is broadly recognized. In rodent studies, for example, activating and de-activating insular cortex was found to have direct and measurable effects on anxiety (Méndez-Ruette et al., 2019). Indeed, inclusion of multiple insular ROIs in the current study was based on results from a meta-analysis of human fMRI studies (Etkin & Wager, 2007) showing frequent implication of the insula in anxiety. Specifically, increased insular activation has been found in many anxiety-related disorders including generalized anxiety disorder, phobias, Obsessive Compulsive Disorder, and Post-Traumatic Stress Disorder (Méndez-Ruette et al., 2019; Xu et al., 2019). Positive correlations between anxiety and functional connectivity of the insula observed in our ASD cohort appears consistent with such fMRI findings, as greater co-activation between insula and other anxiety-network ROIs was associated with greater anxiety. On the other hand, our group-level finding of insular underconnectivity in ASDs appears less readily compatible with findings from non-ASD populations, possibly for reasons discussed in the previous section. Note, however, that stimulus-driven activation in fMRI cannot be equated with correlations of intrinsic activity changes detected in resting state functional connectivity MRI, as implemented in our study.

The roles of the insula as a major hub of the salience network may help explain the prominence of insular ROIs as a critical mechanism related to anxiety in our study. The salience network (SN) is hypothesized to help promote coordination of neural resources by detecting and relaying critically important (or, “salient”) visceral, auditory, vestibular, nociceptive, and thermal input stimuli to facilitate sensory and affective processing and higher-level cognition (Seeley et al., 2007; Uddin et al., 2017). Anxiety is associated with threat-related attentional bias (Bar-Haim, Lamy, Pergamin, Bakermans-Kranenburg, & van IJzendoorn, 2007) which, by definition, includes attending to salient threat stimuli and could therefore be linked to salience-network involvement. In addition to threat-related attentional bias to external stimuli, higher attention towards interoceptive stimuli (bodily sensations, e.g. heart rate) and perception of these as harmful (as in, “my heart racing means something is wrong with it”) has also been suggested as a mechanism by which anxiety develops and persists (Barlow, 2011). As mentioned, the insula receives input related to external stimuli with various levels of salience, but also interoceptive and viscerosensory inputs (from within our own bodies) (Critchley & Harrison, 2013). In other words, there appears to be a neural differentiation between the representation of sensation and perception that follows a posterior-to-anterior progression within the human insula (Craig, 2002). Thus, the insula may play a critical role in translating threat-related attentional biases (both to internal and external stimuli) into anxiety and, in extreme cases, anxiety disorders. However, results of the current study suggest that such links differ between ASD and TC populations, as the relation between anxiety and functional connectivity was different and possibly reversed. In particular, prominent group differences and interactions involving posterior insula could suggest that in adults with ASDs, anxiety is associated with disrupted representation of external or interoceptive stimuli (or both) in this region. This could have translational implications, as treatments developed for non-ASD populations, such as exposure therapy (which includes interoceptive exposures), may not work in similar ways in ASDs.

Relatively few findings involving amygdala

The current study adds to the literature showing atypical amygdalar functional connectivity within the anxiety network in adults with ASDs (Baur, Hänggi, Langer, & Jäncke, 2013; Kleinhans et al., 2016). However, altered amygdalar functional connectivity was less prominent than atypical functional connectivity involving the insula. This unexpected pattern may be explained in several ways. Firstly, the amygdala is a highly diverse structure built of many sub-nuclei with many different roles, from detection of threat to olfaction and averaging the mean BOLD timeseries across all of these sub-nuclei may have masked more nuanced differences related to anxiety as suggested by (Kleinhans et al., 2016). Secondly, higher order functions of the anxiety network related to stimulus salience may be those that are most affected in adults with ASDs.

Limitations

Participants in this study were born between 1950 and 1980, before the first appearance of “Infantile Autism” in the DSM-III (American Psychiatric Association, 1985), when it was believed that the majority of those with autism had IQs below 70. In addition, overall prevalence was estimated at ~0.03%, in contrast to current estimates approaching 2% (Maenner, 2020). Identifying middle-aged adults with ASDs today, particularly those capable of participating in MRI studies, is therefore extremely challenging, as many never received a diagnosis as children or young adults. Sample size of the current study was correspondingly limited, warranting cautious interpretation of findings. Additionally, there is a need for longitudinal research on anxiety in this population to clarify directionality of relationships between neurobiology and behavioral symptoms, and to delve into complex interactions between aging, network reorganization, anxiety, and autism. Although participants were not specifically screened for degenerative disorders none were detected in extensive medical history and the neuropsychological testing conducted. However, given the age-range studied here, it is possible that mild pathological conditions (e.g. mild neurocognitive disorder) were present or incipient in some of the participants, regardless of group, which could have influenced some of the fMRI measurements. We also note here that the majority of participants in the current study had average or above average FSIQ and therefore may be at higher risk for anxiety [(Hollocks et al., 2019; Mingins, Tarver, Waite, Jones, & Surtees, 2021)]). Results may therefore not generalize to individuals with lower cognitive abilities or to non-verbal adults. Additionally, groups were not matched on medication-status (see Supplementary Table S1 for prescription medication breakdown by group), and higher rates of medication usage in the ASD group (indeed partly related to higher anxiety levels) are a potential confound.

Finally, similar to previous research on the neural substrates of anxiety in ASDs (Kleinhans et al., 2016), we used the BAI as a general measure of anxiety symptoms. This measure may not be optimally sensitive to subtle differences associated with distinct anxiety disorders (e.g. social anxiety vs. specific phobia) and/or the presence of state vs. trait anxiety. Alternative measures of anxiety may capture discrete aspects of anxiety disorders (e.g. physiological vs. psychological). Future work incorporating clinical interview assessment measures, some of which have been designed to help disentangle anxiety and ASD symptoms that can appear to overlap (Kerns & Kendall, 2012; Kerns, Renno, Kendall, Wood, & Storch, 2017), is also warranted.. However, given the dearth of fMRI research on anxiety in ASDs and on adults with ASDs, this study is a good first step towards understanding neural substrates of co-occurring anxiety in adults with ASDs.

Conclusions

Our findings show that functional connectivity within an empirically defined anxiety network is atypical in middle-aged adults with ASDs and related to anxiety symptoms, independent of core sociocommunicative deficits. Results prominently implicate bilateral insulae, possibly linked to their role as a part of the salience network. They further suggest that the neural substrates of anxiety in ASDs may differ from those in TC adults, possibly indicating specific mechanisms of biological vulnerability to anxiety in adults with ASDs. As this may have implications with respect to cognitive-behavioral and biological interventions (e.g. SSRI medication), further research is needed, given the dearth of studies on middle-aged and older adults with ASDs.

Supplementary Material

supinfo

Acknowledgements:

Our sincere thanks to our participants and their families for sharing their time with us.

Funding:

This study was supported by the National Institutes of Health [grant numbers NIH R01-MH103494 (awarded to Ralph-Axel Müller, Ph.D.)], by the National Institute of General Medical Sciences [grant number NIH 5T34GM008303 (awarded to Ryan Tung)], and by Autism Speaks [Weatherstone fellowship #10609 (awarded to Maya Reiter M.S)]. The presented content (including study design, data analysis and written manuscript) is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

Footnotes

Disclosures:

Author R. Tung has nothing to disclose.

Author M. A Reiter has nothing to disclose.

Author A. Linke has nothing to disclose.

Author J.S. Kohli has nothing to disclose.

Author M. K. Kinnear has nothing to disclose.

Author R.-A. Müller has nothing to disclose.

Author R. A. Carper has nothing to disclose.

5. References

  1. Amaral DG, & Corbett BA (2003). The amygdala, autism and anxiety. Novartis Foundation Symposium, 251, 177–187; -197, 281–297. [PubMed] [Google Scholar]
  2. American Psychiatric Association. (1985). Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition (3rd edition). Washington, D.C.: The American Psychiatric Association. [Google Scholar]
  3. American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders, 5th ed. American Psychiatric Association, Washington, DC. [Google Scholar]
  4. Bar-Haim Y, Lamy D, Pergamin L, Bakermans-Kranenburg MJ, & van IJzendoorn MH (2007). Threat-related attentional bias in anxious and nonanxious individuals: A meta-analytic study. Psychological Bulletin, 133(1), 1–24. 10.1037/0033-2909.133.1.1 [DOI] [PubMed] [Google Scholar]
  5. Barlow DH (2011). Chapter 1: Panic disorder and agoraphobia. In Clinical handbook of psychological disorders, a step by step treatment manual (pp. 7–8). The Guilford Press. [Google Scholar]
  6. Baur V, Hänggi J, Langer N, & Jäncke L (2013). Resting-State Functional and Structural Connectivity Within an Insula–Amygdala Route Specifically Index State and Trait Anxiety. Biological Psychiatry, 73(1), 85–92. 10.1016/j.biopsych.2012.06.003 [DOI] [PubMed] [Google Scholar]
  7. Behzadi Y, Restom K, Liau J, & Liu TT (2007). A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. NeuroImage, 37(1), 90–101. 10.1016/j.neuroimage.2007.04.042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bijsterbosch J, Harrison S, Duff E, Alfaro-Almagro F, Woolrich M, & Smith S (2017). Investigations into within- and between-subject resting-state amplitude variations. NeuroImage, 159, 57–69. 10.1016/j.neuroimage.2017.07.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Buck TR, Viskochil J, Farley M, Coon H, McMahon WM, Morgan J, & Bilder DA (2014). Psychiatric Comorbidity and Medication Use in Adults with Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 44(12), 3063–3071. 10.1007/s10803-014-2170-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Byers AL, Yaffe K, Covinsky KE, Friedman MB, & Bruce ML (2010). High Occurrence of Mood and Anxiety Disorders among Older Adults: The National Comorbidity Survey Replication. Archives of General Psychiatry, 67(5), 489–496. 10.1001/archgenpsychiatry.2010.35 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Craig AD (2002). How do you feel? Interoception: the sense of the physiological condition of the body. Nature Reviews Neuroscience, 3(8), 655-. Gale Academic OneFile. Retrieved from Gale Academic OneFile. [DOI] [PubMed] [Google Scholar]
  12. Crawford J, Cayley C, Lovibond PF, Wilson PH, & Hartley C (2011). Percentile Norms and Accompanying Interval Estimates from an Australian General Adult Population Sample for Self-Report Mood Scales (BAI, BDI, CRSD, CES-D, DASS, DASS-21, STAI-X, STAI-Y, SRDS, and SRAS). Australian Psychologist, 46(1), 3–14. 10.1111/j.1742-9544.2010.00003.x [DOI] [Google Scholar]
  13. Critchley HD, & Harrison NA (2013). Visceral Influences on Brain and Behavior. Neuron, 77(4), 624–638. 10.1016/j.neuron.2013.02.008 [DOI] [PubMed] [Google Scholar]
  14. Cvejic RC, Arnold SRC, Foley K-R, & Trollor JN (2018). Neuropsychiatric profile and psychotropic medication use in adults with autism spectrum disorder: Results from the Australian Longitudinal Study of Adults with Autism. BJPsych Open, 4(6), 461–466. 10.1192/bjo.2018.64 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Davis TE, Hess JA, Moree BN, Fodstad JC, Dempsey T, Jenkins WS, & Matson JL (2011). Anxiety symptoms across the lifespan in people diagnosed with Autistic Disorder. Research in Autism Spectrum Disorders, 5(1), 112–118. 10.1016/j.rasd.2010.02.006 [DOI] [Google Scholar]
  16. de Carvalho MR, Dias GP, Cosci F, de-Melo-Neto VL, Bevilaqua MC do N, Gardino PF, & Nardi AE (2010). Current findings of fMRI in panic disorder: Contributions for the fear neurocircuitry and CBT effects. Expert Review of Neurotherapeutics, 10(2), 291–303. 10.1586/ern.09.161 [DOI] [PubMed] [Google Scholar]
  17. Etkin A, & Wager TD (2007). Functional Neuroimaging of Anxiety: A Meta-Analysis of Emotional Processing in PTSD, Social Anxiety Disorder, and Specific Phobia. The American Journal of Psychiatry, 164(10), 1476–1488. 10.1176/appi.ajp.2007.07030504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fair DA, Miranda-Dominguez O, Snyder AZ, Perrone A, Earl EA, Van AN, … Dosenbach NUF (2020). Correction of respiratory artifacts in MRI head motion estimates. NeuroImage, 208, 116400. 10.1016/j.neuroimage.2019.116400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Geerligs L, Renken RJ, Saliasi E, Maurits NM, & Lorist MM (2015). A Brain-Wide Study of Age-Related Changes in Functional Connectivity. Cerebral Cortex, 25(7), 1987–1999. 10.1093/cercor/bhu012 [DOI] [PubMed] [Google Scholar]
  20. Gillott A, & Standen PJ (2007). Levels of anxiety and sources of stress in adults with autism. Journal of Intellectual Disabilities, 11(4), 359–370. 10.1177/1744629507083585 [DOI] [PubMed] [Google Scholar]
  21. Gotham K, Brunwasser SM, & Lord C (2015). Depressive and Anxiety Symptom Trajectories From School-Age Through Young Adulthood in Samples With Autism Spectrum Disorder and Developmental Delay. Journal of the American Academy of Child and Adolescent Psychiatry, 54(5), 369–376.e3. 10.1016/j.jaac.2015.02.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. He X, Qin W, Liu Y, Zhang X, Duan Y, Song J, … Yu C (2014). Abnormal salience network in normal aging and in amnestic mild cognitive impairment and Alzheimer’s disease. Human Brain Mapping, 35(7), 3446–3464. 10.1002/hbm.22414 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hennessey T, Andari E, & Rainnie DG (2018). RDoC-based categorization of amygdala functions and its implications in autism. Neuroscience and Biobehavioral Reviews, 90, 115–129. 10.1016/j.neubiorev.2018.04.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Herrington JD, Maddox BB, Kerns CM, Rump K, Worley JA, Bush JC, … Miller JS (2017). Amygdala Volume Differences in Autism Spectrum Disorder Are Related to Anxiety. Journal of Autism and Developmental Disorders, 47(12), 3682–3691. 10.1007/s10803-017-3206-1 [DOI] [PubMed] [Google Scholar]
  25. Hollocks MJ, Lerh JW, Magiati I, Meiser-Stedman R, & Brugha TS (2019). Anxiety and depression in adults with autism spectrum disorder: A systematic review and meta-analysis. Psychological Medicine, 49(4), 559–572. 10.1017/S0033291718002283 [DOI] [PubMed] [Google Scholar]
  26. James GM, Baldinger-Melich P, Philippe C, Kranz GS, Vanicek T, Hahn A, … Lanzenberger R (2017). Effects of Selective Serotonin Reuptake Inhibitors on Interregional Relation of Serotonin Transporter Availability in Major Depression. Frontiers in Human Neuroscience, 11. 10.3389/fnhum.2017.00048 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Juranek J, Filipek PA, Berenji GR, Modahl C, Osann K, & Spence MA (2006). Association Between Amygdala Volume and Anxiety Level: Magnetic Resonance Imaging (MRI) Study in Autistic Children. Journal of Child Neurology, 21(12), 1051–1058. 10.1177/7010.2006.00237 [DOI] [PubMed] [Google Scholar]
  28. Kasper L, Bollmann S, Diaconescu AO, Hutton C, Heinzle J, Iglesias S, … Stephan KE (2017). The PhysIO Toolbox for Modeling Physiological Noise in fMRI Data. Journal of Neuroscience Methods, 276, 56–72. 10.1016/j.jneumeth.2016.10.019 [DOI] [PubMed] [Google Scholar]
  29. Kerns, Connor M, & Kendall PC (2012). The Presentation and Classification of Anxiety in Autism Spectrum Disorder. Clinical Psychology: Science and Practice, 19(4), 323–347. 10.1111/cpsp.12009 [DOI] [Google Scholar]
  30. Kerns, Connor M, Renno P, Kendall PC, Wood JJ, & Storch EA (2017). Anxiety Disorders Interview Schedule–Autism Addendum: Reliability and Validity in Children With Autism Spectrum Disorder. Journal of Clinical Child & Adolescent Psychology, 46(1), 88–100. 10.1080/15374416.2016.1233501 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kleinhans NM, Reiter MA, Neuhaus E, Pauley G, Martin N, Dager S, & Estes A (2016). Subregional differences in intrinsic amygdala hyper and hypo connectivity in autism spectrum disorder. Autism Research: Official Journal of the International Society for Autism Research, 9(7), 760–772. 10.1002/aur.1589 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kleinhans NM, Richards T, Weaver K, Johnson LC, Greenson J, Dawson G, & Aylward E (2010). Association between amygdala response to emotional faces and social anxiety in autism spectrum disorders. Neuropsychologia, 48(12), 3665–3670. 10.1016/j.neuropsychologia.2010.07.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kolesar TA, Bilevicius E, Wilson AD, & Kornelsen J (2019). Systematic review and meta-analyses of neural structural and functional differences in generalized anxiety disorder and healthy controls using magnetic resonance imaging. NeuroImage : Clinical, 24. 10.1016/j.nicl.2019.102016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lenze EJ, & Wetherell JL (2011). A lifespan view of anxiety disorders. Dialogues in Clinical Neuroscience, 13(4), 381–399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lever AG, & Geurts HM (2016). Psychiatric Co-occurring Symptoms and Disorders in Young, Middle-Aged, and Older Adults with Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 46(6), 1916–1930. 10.1007/s10803-016-2722-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lord C, Risi S, Lambrecht L, Cook EH, Leventhal BL, DiLavore PC, … Rutter M (2014). The Autism Diagnostic Observation Schedule–Generic: A Standard Measure of Social and Communication Deficits Associated with the Spectrum of Autism. 19. [PubMed] [Google Scholar]
  37. Maenner MJ (2020). Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years—Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2016. MMWR. Surveillance Summaries, 69. 10.15585/mmwr.ss6904a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Magiati I, Tay XW, & Howlin P (2014). Cognitive, language, social and behavioural outcomes in adults with autism spectrum disorders: A systematic review of longitudinal follow-up studies in adulthood. Clinical Psychology Review, 34(1), 73–86. 10.1016/j.cpr.2013.11.002 [DOI] [PubMed] [Google Scholar]
  39. Méndez-Ruette M, Linsambarth S, Moraga-Amaro R, Quintana-Donoso D, Méndez L, Tamburini G, … Stehberg J (2019). The Role of the Rodent Insula in Anxiety. Frontiers in Physiology, 10. 10.3389/fphys.2019.00330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mikita N, Simonoff E, Pine DS, Goodman R, Artiges E, Banaschewski T, … Stringaris A (2016). Disentangling the autism−anxiety overlap: FMRI of reward processing in a community-based longitudinal study. Translational Psychiatry, 6(6), e845. 10.1038/tp.2016.107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mingins JE, Tarver J, Waite J, Jones C, & Surtees AD (2021). Anxiety and intellectual functioning in autistic children: A systematic review and meta-analysis. Autism, 25(1), 18–32. 10.1177/1362361320953253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Mochcovitch MD, da Rocha Freire RC, Garcia RF, & Nardi AE (2014). A systematic review of fMRI studies in generalized anxiety disorder: Evaluating its neural and cognitive basis. Journal of Affective Disorders, 167, 336–342. 10.1016/j.jad.2014.06.041 [DOI] [PubMed] [Google Scholar]
  43. Nakao T, Okada K, & Kanba S (2014). Neurobiological model of obsessive-compulsive disorder: Evidence from recent neuropsychological and neuroimaging findings. Psychiatry and Clinical Neurosciences, 68(8), 587–605. 10.1111/pcn.12195 [DOI] [PubMed] [Google Scholar]
  44. Nimmo-Smith V, Heuvelman H, Dalman C, Lundberg M, Idring S, Carpenter P, … Rai D (2020). Anxiety Disorders in Adults with Autism Spectrum Disorder: A Population-Based Study. Journal of Autism and Developmental Disorders, 50(1), 308–318. 10.1007/s10803-019-04234-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Onoda K, Ishihara M, & Yamaguchi S (2012). Decreased Functional Connectivity by Aging Is Associated with Cognitive Decline. Journal of Cognitive Neuroscience, 24(11), 2186–2198. 10.1162/jocn_a_00269 [DOI] [PubMed] [Google Scholar]
  46. Seeley WW, Menon V, Schatzberg AF, Keller J, Glover GH, Kenna H, … Greicius MD (2007). Dissociable Intrinsic Connectivity Networks for Salience Processing and Executive Control. The Journal of Neuroscience, 27(9), 2349–2356. 10.1523/JNEUROSCI.5587-06.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, Johansen-Berg H, … Matthews PM (2004). Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage, 23 Suppl 1, S208–219. 10.1016/j.neuroimage.2004.07.051 [DOI] [PubMed] [Google Scholar]
  48. Steer RA, & Beck AT (1997). Beck Anxiety Inventory. Evaluating Stress: A Book of Resources, 23–40. [Google Scholar]
  49. Storey JD (2002). A direct approach to false discovery rates. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 64(3), 479–498. 10.1111/1467-9868.00346 [DOI] [Google Scholar]
  50. Tomasi D, & Volkow ND (2012). Aging and functional brain networks. Molecular Psychiatry, 17(5), 549–558. 10.1038/mp.2011.81 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Top DN Jr., Stephenson KG, Doxey CR, Crowley MJ, Kirwan CB, & South M (2016). Atypical Amygdala Response to Fear Conditioning in Autism Spectrum Disorder. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 1(4), 308–315. 10.1016/j.bpsc.2016.01.008 [DOI] [PubMed] [Google Scholar]
  52. Uddin LQ, Nomi JS, Hébert-Seropian B, Ghaziri J, & Boucher O (2017). Structure and Function of the Human Insula: Journal of Clinical Neurophysiology, 34(4), 300–306. 10.1097/WNP.0000000000000377 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Uljarević M, Hedley D, Rose-Foley K, Magiati I, Cai RY, Dissanayake C, … Trollor J (2020). Anxiety and Depression from Adolescence to Old Age in Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 50(9), 3155–3165. 10.1007/s10803-019-04084-z [DOI] [PubMed] [Google Scholar]
  54. van Heijst BF, & Geurts HM (2015). Quality of life in autism across the lifespan: A meta-analysis. Autism, 19(2), 158–167. 10.1177/1362361313517053 [DOI] [PubMed] [Google Scholar]
  55. van Steensel FJA, Bögels SM, & Dirksen CD (2012). Anxiety and Quality of Life: Clinically Anxious Children With and Without Autism Spectrum Disorders Compared. Journal of Clinical Child & Adolescent Psychology, 41(6), 731–738. 10.1080/15374416.2012.698725 [DOI] [PubMed] [Google Scholar]
  56. Varangis E, Habeck CG, Razlighi QR, & Stern Y (2019). The Effect of Aging on Resting State Connectivity of Predefined Networks in the Brain. Frontiers in Aging Neuroscience, 11. 10.3389/fnagi.2019.00234 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Weber-Goericke F, & Muehlhan M (2019). A quantitative meta-analysis of fMRI studies investigating emotional processing in excessive worriers: Application of activation likelihood estimation analysis. Journal of Affective Disorders, 243, 348–359. 10.1016/j.jad.2018.09.049 [DOI] [PubMed] [Google Scholar]
  58. Wechsler D (2011). Wechsler Abbreviated Scale of Intelligence, Second Edition (WASI-II). [Google Scholar]
  59. White N, Roddey C, Shankaranarayanan A, Han E, Rettmann D, Santos J, … Dale A (2010). PROMO: Real-time prospective motion correction in MRI using image-based tracking. Magnetic Resonance in Medicine, 63(1), 91–105. 10.1002/mrm.22176 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Xu J, Van Dam NT, Feng C, Luo Y, Ai H, Gu R, & Xu P (2019). Anxious brain networks: A coordinate-based activation likelihood estimation meta-analysis of resting-state functional connectivity studies in anxiety. Neuroscience & Biobehavioral Reviews, 96, 21–30. 10.1016/j.neubiorev.2018.11.005 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supinfo
NIHMS1722530-supplement-supinfo.docx (4.7MB, docx)

RESOURCES