- Email: [email protected]
Resting state functional connectivity correlates of emotional awareness
Accepted Manuscript Resting state functional connectivity correlates of emotional awareness Ryan Smith, Anna Alkozei, Jennifer Bao, Courtney Smith, Richard D. Lane, William D.S. Killgore PII:
S1053-8119(17)30612-2
DOI:
10.1016/j.neuroimage.2017.07.044
Reference:
YNIMG 14209
To appear in:
NeuroImage
Received Date: 24 January 2017 Revised Date:
16 June 2017
Accepted Date: 19 July 2017
Please cite this article as: Smith, R., Alkozei, A., Bao, J., Smith, C., Lane, R.D., Killgore, W.D.S., Resting state functional connectivity correlates of emotional awareness, NeuroImage (2017), doi: 10.1016/ j.neuroimage.2017.07.044. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Resting State Functional Connectivity Correlates of Emotional Awareness Ryan Smith1, Anna Alkozei1, Jennifer Bao1, Courtney Smith1, Richard D. Lane1, William D. S. Killgore1
Department of Psychiatry, University of Arizona, 1501 N. Campbell Ave, Tucson, AZ 85724-5002,
RI PT
1
Corresponding author at:
TE D
M AN U
SC
United States
EP
University of Arizona, Department of Psychiatry, 1501 N., Campbell Ave., PO Box 245002, Room 7304B, Tucson, AZ 85724-5002, United States. Tel.: +1 602 501 4168; fax: +1 520 626 6050.
AC C
E-mail address: [email protected] (Ryan Smith).
ACCEPTED MANUSCRIPT 2 Abstract Multiple neuroimaging studies have now linked emotional awareness (EA), as measured by the Levels of Emotional Awareness Scale (LEAS), with activation in regions of neural networks associated with both
RI PT
conceptualization (i.e., default mode network [DMN] regions) and interoception (i.e., salience network [SN] regions) – consistent with the definition of EA as one’s ability to appropriately recognize,
conceptualize, and articulate the emotions of self and other in fine-grained, differentiated ways. However, no study has yet tested the hypothesis that greater LEAS scores are associated with greater resting state
SC
functional connectivity (FC) within these networks. Twenty-six adults (13 female) underwent resting state functional magnetic resonance imaging, and also completed the LEAS. Using pre-defined functional ROIs
M AN U
from the DMN and SN, we observed that LEAS scores were significantly positively correlated with FC between several regions of both of these networks, even when controlling for differences in general intelligence (IQ). These results suggest that higher EA may be associated with more efficient information exchange between brain regions involved in both interoception- and conceptualization-based processing, which could plausibly contribute to more differentiated bodily feelings and more fine-grained
TE D
conceptualization of those feelings.
Keywords: Emotional Awareness; Levels of Emotional Awareness Scale (LEAS); Default Mode Network;
AC C
EP
Salience Network; Conceptualization; Interoception
ACCEPTED MANUSCRIPT 3 Introduction
It is a commonplace clinical observation that there are substantial individual differences in people’s awareness of their own emotions. Specifically, some individuals appear to have a greater ability than others
RI PT
to understand, and clearly articulate, the emotions they are feeling. One well-established and validated
measure of these individual differences in emotional awareness (EA) is the Levels of Emotional Awareness Scale (LEAS) (Lane et al., 1990; Lane and Schwartz, 1987). As measured by the LEAS, higher EA has been found to correlate positively with many adaptive traits/abilities, including self-reported impulse
SC
control (Bréjard et al., 2012), openness to feelings (Lane et al., 1990), emotion recognition ability (Lane et al., 2000, 1996), empathy (Barchard and Hakstian, 2004), and a stable sense of well-being independent of
M AN U
current mood (Ciarrochi et al., 2003). In contrast, lower LEAS scores have been associated with several maladaptive clinical phenomena, such as essential hypertension (Consoli et al., 2010), eating disorders (Bydlowski et al., 2005), post-traumatic stress disorder (Frewen et al., 2008), schizophrenia (Baslet et al., 2009), depression (Berthoz et al., 2000; Donges et al., 2005), borderline personality disorder (Levine et al., 1997), somatoform disorders (Subic-Wrana et al., 2005), a “disorganized attachment style” (Subic-Wrana
TE D
et al., 2007), impaired insight in the context of cocaine abuse (Moeller et al., 2014), and greater pain in patients with irritable bowel syndrome (IBS) (Lackner, 2005). Thus, EA appears to represent an important individual difference variable associated with both emotional and physical health.
EP
The neural basis of EA has also begun to receive experimental investigation. For example, two different task-based functional neuroimaging studies have demonstrated that higher LEAS scores are associated with
AC C
greater activity in the dorsal anterior cingulate cortex (dACC) (Lane et al., 1998a; McRae et al., 2008) – a region of the “salience network” (SN; Barrett & Satpute, 2013) implicated in autonomic regulation (Critchley et al., 2003), attention to one’s bodily states (Farb et al., 2013), and facilitating the influence of bodily feelings on action selection (Critchley, 2005; Medford and Critchley, 2010). Another study has also shown that, during recall of life-threatening experiences, healthy subjects showed greater activity in the rostral ACC (rACC) as a function of greater scores on the LEAS (Frewen et al., 2008). The rACC, and adjacent parts of the medial prefrontal cortex (MPFC), are known to comprise a significant hub of the “default mode network” (DMN) (Barrett and Satpute, 2013; Buckner et al., 2008; Li et al., 2014; Raichle et
ACCEPTED MANUSCRIPT 4 al., 2001), which refers to a set of brain regions whose activity is highly correlated at rest and which are thought to play an important role in the process of conceptualization. In combination with the results of other studies of the rACC/MPFC (Kalisch et al., 2006; Peelen et al., 2010; Roy et al., 2012; R Smith et al.,
RI PT
2014b), we have previously suggested that this region, in conjunction with the rest of the DMN, may play an important role in assigning conceptual significance to bodily feelings (Lane et al., 2015; Smith and
Lane, 2015) – and therefore allow such feelings to be understood in explicit emotional terms (e.g., sadness, fear, etc.). This suggestion is also consistent with another study (Tavares et al., 2011), which showed that,
SC
while subjects viewed simple animated scenarios with social/emotional content, higher LEAS scores
predicted greater neural activity within abstract semantic processing regions (i.e., left anterior temporal
M AN U
cortex), whereas lower LEAS scores predicted more concrete action-oriented brain activation (i.e., in premotor cortex).
While the studies described above have examined the association between EA and task-related neural responses, it is notable that no study to date has yet examined the relation between LEAS scores and resting state measures of functional connectivity (FC) in the DMN or SN. Given that DMN regions have
TE D
themselves been largely characterized using resting state connectivity measures (Buckner et al., 2008), and given the theoretical/empirical work described above supporting a link between EA and DMN regions/functions, there are strong reasons to hypothesize that LEAS scores should predict individual differences in DMN connectivity at rest. Given that EA also involves recognition/articulation of
EP
differentiated bodily feeling states, and that the SN is involved in representing bodily percepts (and using them to guide cognition and action selection), there is also reason to predict LEAS scores would be
AC C
associated with better FC between SN regions. In the present study, we therefore examined the correlations between LEAS scores and individual differences in resting state FC with the hypothesis that higher LEAS scores would be associated with stronger positive connectivity between DMN regions and SN regions, respectively.
Materials and Methods
Participants
ACCEPTED MANUSCRIPT 5 Twenty-six adults (13 female; mean age = 23.12 +/- 4.03) were recruited from the general population via flyers and internet advertisements to participate in the present study. Participants did not have any history of psychiatric or neurological disorders (assessed via a phone screen questionnaire based on criteria within
RI PT
the Diagnostic and Statistical Manual for Mental Disorders, 4th addition; DSM-IV-TR), and all provided written informed consent prior to participation. All participants received a nominal financial compensation for participation. The research protocol of the present study was also reviewed and approved by the Institutional Review Board of the University of Arizona.
SC
Procedure
M AN U
Upon completing the informed consent process, participants were taken to the magnetic resonance imaging (MRI) scanner at the University of Arizona where they underwent a resting state functional scan (see Neuroimaging Methods below). After completing the resting state functional scan, participants were escorted back to the lab, seated at a laptop, and asked to complete an on-line version of the LEAS (www.eleastest.net) that makes use of a validated automatic scoring program (Barchard et al., 2010).
TE D
LEAS. The LEAS presents participants with 2-4 sentence descriptions of 20 social situations, where each situation includes 2 people. The situation descriptions are designed to elicit four types of emotion (sadness, happiness, anger, and fear) at 5 levels of complexity. One situation is presented on each electronically presented page, followed by two questions: “How would you feel?” and “How would the other person
EP
feel?” Separate response boxes are provided for typing in the answers to each question. Participants are instructed to type their responses into these boxes, and they are asked to use as much or as little space as
AC C
needed to answer. The only rule given is that they must use the word “feel” in their responses.
Scores reflecting EA level are assigned based on the words participants provide in their responses. The lowest scores (Level 0) are given to words that do not refer to feelings. Level 1 scores are given to words that refer to physiological sensations (e.g., “tired”), whereas level 2 scores are given to words referring to feeling-related actions (e.g., “punching”) or simple valence discriminations (e.g., “bad,” “good”) that have inherent avoidance- or approach-related content. Level 3 scores are assigned to words referring to single emotion concepts (e.g., “happy,” “sad”). Level 4 scores are given when at least 2 words from level 3 are
ACCEPTED MANUSCRIPT 6 used (i.e., when they convey greater emotional differentiation than either word alone). For each item, the self- and other-related responses are scored separately (i.e., with a value of 0-4). A “total” score is also given for each of the 20 LEAS items; this score reflects the higher of the self- and other-related scores,
RI PT
unless a score of 4 is given for both. In this case, a total score of 5 is given for the item, so long as the selfand other-related responses are sufficiently differentiable (for more detail, see Lane et al., 1990).
General Intelligence. Intelligence quotient (IQ) was measured with the two-subtest form (FSIQ-2) of the Wechsler Abbreviated Scale of Intelligence – Second Edition (WASI-II; Pearson Assessment, Inc., San
SC
Antonio, TX; Wechsler, 2011). This was done in order to control for general intelligence when examining
M AN U
FC correlates of LEAS scores.
Neuroimaging Methods
Neuroimaging was performed within a 3T Siemens Skyra scanner (Siemens, Erlangen, Germany) with a 32-channel head coil. T1-weighted structural images (3D MPRAGE) were acquired (TR/TE/flip angle = 2.1 s / 2.33 ms/ 12 degree) covering 176 sagittal slices (256 x 256) with a slice thickness of 1 mm (voxel
TE D
size = 1 x 1 x 1). Functional T2*-weighted scans were acquired over 32 transverse slices (2.5 mm thickness; matrix: 88x84). Each volume was collected with an interleaved sequence (TR/TE/flip angle = 2 s/ 25 ms/ 90 degree). The voxel size of the T2* sequence was 2.5 x 2.5 x 3.5 mm (i.e., with a 40% slice
240 mm.
EP
gap, allowing collection of 300 volumes within a 10-minute acquisition time). The field of view (FOV) was
AC C
Resting-state preprocessing
The publicly available CONN functional connectivity toolbox (version 16.a; https://www.nitrc.org/projects/conn), in conjunction with SPM12 (Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion.ucl.ac.uk/spm), was used to perform all preprocessing steps (using CONN’s default preprocessing pipeline), as well as subsequent statistical analyses, on all collected MRI scans. In this preprocessing pipeline, raw functional images were slice-time corrected, realigned (motion corrected), unwarped, and coregistered to each subject’s MPRAGE image in accordance with standard algorithms. Images were then normalized to Montreal Neurological Institute (MNI) coordinate
ACCEPTED MANUSCRIPT 7 space, spatially smoothed (8 mm full-width at half maximum), and resliced to 2 x 2 x 2 mm voxels. The Artifact Detection Tool (ART; http://www.nitrc.org/projects/artifact_detect/) was also used to regress out scans as nuisance covariates in the first-level analysis exceeding 3 SD in mean global intensity and scan-to-
RI PT
scan motion that exceeded 0.5 mm. These were added in addition to covariates for the 6 rotation/translation movement parameters.
Functional Connectivity Analysis
SC
Using a standard seed-driven approach, FC analyses were performed using the default FC processing pipeline in the CONN toolbox (for details, see Whitfield-Gabrieli & Nieto-Castanon, 2012). In this
M AN U
processing pipeline, physiological and other spurious sources of noise were estimated with the aCompcor method (Behzadi et al., 2007; Chai et al., 2012; Whitfield-Gabrieli et al., 2009); they were then removed together with the movement- and artifact-related covariates mentioned above. The residual BOLD timeseries was then band-pass filtered (.008Hz–.09 Hz). Every participant's structural image was segmented into gray matter, white matter, and cerebral spinal fluid using SPM12. White matter and cerebral spinal fluid noise ROIs were removed through regression. The ROIs we used were taken from a freely available
TE D
atlas of regions defined by correlated activation patterns (http://findlab.stanford.edu/functional_ROIs.html). This atlas includes 90 ROIs, grouped into several networks: The anterior and posterior SN, the dorsal and ventral DMN, the left and right executive control network, as well as the auditory network, basal ganglia
EP
network, higher visual network, language network, sensorimotor network, primary visual network, visuospatial network, and precuneus network (for details, see Shirer, Ryali, Rykhlevskaia, Menon, &
AC C
Greicius, 2012). However, we only used the dorsal DMN and anterior/posterior SN ROIs in our analyses.
To produce first-level correlation maps, for each participant the residual BOLD time course was then extracted from each DMN and SN ROI and Pearson's correlation coefficients were computed between a priori selected seed ROI time courses (i.e., 2 seed ROIs from the DMN and 5 seed ROIs from the SN; see the paragraph immediately below for further explanation) and the time courses of all other included ROIs. The resulting correlation coefficients (i.e., one for each seed-target ROI pair in each participant) were then Fisher transformed into ‘Z’ scores to increase normality and thus improve the subsequent second-level General Linear Model analyses.
ACCEPTED MANUSCRIPT 8 At the second level, we then performed two analyses based on our a priori hypotheses (i.e., by compiling/examining the Fisher-transformed Z scores from each participant described above). The first analysis tested the hypothesis that the ROIs within the dorsal DMN would show higher connectivity in
RI PT
individuals with higher LEAS scores. The dorsal DMN includes 9 ROIs (see Figure 1): the MPFC (this ROI also covers parts of the ACC and orbitofrontal cortex [OFC]), the posterior cingulate (PCC; this ROI also covers parts of the precuneus), left hippocampus, a bilateral region of the thalamus, right hippocampus, a region of the right superior frontal gyrus, the mid-cingulate, left angular gyrus, and the right angular
SC
gyrus. For this analysis we chose to use the MPFC ROI and the PCC ROI as seeds, as these are recognized as central hubs within this network and have also been most strongly linked to emotional
M AN U
attention/awareness in previous studies (e.g., Frewen et al., 2008; Gusnard et al., 2001; Lane et al., 2015; R Smith et al., 2014a, 2014b; Smith and Lane, 2015). Connectivity was tested between these two seed ROIs and all other dorsal DMN ROIs listed above. The second analysis tested the hypothesis that the ROIs within the SN (both anterior and posterior network regions together) would have higher connectivity in individuals with higher LEAS scores. The anterior portion of the SN included 7 ROIs (see Figure 1): the left and right anterior insula (AI), the dACC (this ROI also covers parts of the dorsal MPFC and
TE D
supplementary motor area; SMA), regions of the left and right middle frontal gyrus, and regions of the left and right cerebellum. The posterior portion of the SN included 12 ROIs (see Figure 1), including the left and right posterior insula (PI), the left and right supramarginal gyrus (SMG; this ROI also covers parts of
EP
the inferior parietal gyrus [IPG]), regions of the left and right precuneus, regions of the left and right cerebellum, regions of the left and right thalamus, a region of the left middle frontal gyrus, and a region
AC C
within the right mid-cingulate (Shirer, Ryali, Rykhlevskaia, Menon, & Greicius, 2012). As seeds in this analysis, we used the left and right AI, left and right PI, and the dACC ROIs, as the insula and dACC have been most strongly linked to emotional awareness/attention in previous studies (e.g., Craig, 2009; Lane et al., 1998a; McRae et al., 2008; Medford and Critchley, 2010; R Smith et al., 2014a; Smith and Lane, 2015). Connectivity was tested between these five seed ROIs and all other SN ROIs. For these analyses we used an FDR-corrected threshold of p < .05. This correction was applied at the analysis-level, after testing for connectivity between all seed-target ROI pairs in a given network. In these analyses we also included age, gender, and FSIQ-2 scores as covariates of no interest. This was done to ensure that our results only
ACCEPTED MANUSCRIPT 9 revealed FC differences that were explained uniquely by EA, and not attributable to differences in general intelligence, age, or gender (e.g., as suggested by previous work; see Barrett, Lane, Sechrest, & Schwartz, 2000; Lane, Quinlan, Schwartz, Walker, & Zeitlin, 1990b; Lane, Sechrest, & Riedel, 1998; Roberton,
RI PT
Daffern, & Bucks, 2013).
In a final exploratory analysis, we also examined connectivity between each of the DMN and SN hub
regions used as seed regions in the above analyses. This between-network analysis was done to explore
whether differences in EA might also be associated with differences in how the DMN and SN functionally
SC
interact. This analysis used the same covariates and statistical thresholds stated above, and examined
whether LEAS scores were significantly associated with connectivity values between the following ROIs
M AN U
(i.e., all used as both seeds and targets): MPFC, PCC, dACC, left AI, right AI, left PI, and right PI.
Results
Behavioral Results
LEAS scores were as follows: TOTAL = 74.0 (+9.8)%, SELF = 63.0 (+8.7)%, OTHER = 58.5 (+10.7)%.
TE D
Females had numerically higher LEAS TOTAL scores than males (means: females = 76.8 + 10.6; males = 71.2 + 8.4), but this difference was non-significant (t = 1.5, p = .147). The correlation between age and
EP
LEAS scores was negative and non-significant (r = -.315, p = .12).
WASI-II FSIQ-2 scores had a mean of 115.23 (+11.7). LEAS TOTAL scores were significantly positively
AC C
correlated with FSIQ-2 scores (r = .619, p = .001).1
Default Mode Network Analysis
With higher LEAS scores, the MPFC ROI showed greater FC with the PCC, the right hippocampus, and with the left angular gyrus (See Table 1.1; Figures 2 and 3), even after accounting for covariates of age, gender, and FSIQ. With higher LEAS scores, the PCC ROI showed greater FC with the MPFC, the mid-
1 The LEAS scores and FSIQ-2 scores from this data set have previously been published in conjunction with task-related imaging data (Smith et al., in press.). Their relation to resting state imaging data, however, is novel to the present manuscript.
ACCEPTED MANUSCRIPT 10 cingulate, the left and right hippocampus, and with bilateral thalamus (See Table 1.2; Figures 2 and 3). As also reported in Table 1, the mean connectivity values for all but one of these seed-target pairs were significantly greater than zero. The only exception was mean connectivity between the PCC and the
RI PT
bilateral thalamus, which was numerically positive but not significantly different from zero.
Salience Network Analysis
With higher LEAS scores, the dACC ROI showed greater FC with two ROIs in the left cerebellum (i.e.,
SC
one from the anterior and one from the posterior SN; See Table 2.1; Figures 2 and 3). Higher LEAS scores were associated with greater FC between the right AI ROI and the left precuneus and left SMG (i.e., both
M AN U
from the posterior SN; See Table 2.2). With higher LEAS scores, the left PI ROI showed greater FC with the left and right cerebellum ROIs from the anterior SN (See Table 2.3; Figures 2 and 3) 2. The left AI and right PI seeds did not show greater FC with any other SN ROIs with increasing LEAS scores. As also reported in Table 2, the mean connectivity values for all but one of these seed-target pairs were significantly greater than zero. The only exception was mean connectivity between the left PI and the right
Between Network Analysis
TE D
cerebellum, which was not significantly different from zero.
No statistically significant associations were observed between LEAS scores and FC between any of the
EP
seed and target ROIs examined in this analysis (i.e., between the two DMN ROIs and the five SN ROIs used in the other analyses described above).
AC C
Discussion
In this study we tested the hypotheses that greater EA would be associated with stronger resting state FC between regions of pre-established cortical networks subserving 1) conceptualization (i.e., the DMN) and 2) processing of afferent bodily signals (i.e., the SN) (Barrett and Satpute, 2013). In support of the first hypothesis, we found that higher EA (as measured by LEAS scores) predicted stronger FC between ROIs in
2
Tables 1-2 only report significant results. For the full results of each of the analyses, including nonsignificant results, the reader is referred to the supplementary materials.
ACCEPTED MANUSCRIPT 11 several DMN regions, including between the MPFC and the PCC, right hippocampus, and left angular gyrus, as well as between the PCC and the MPFC, the left and right hippocampus, the mid-cingulate, and the thalamus. In support of the second hypothesis, we found that higher EA predicted stronger FC between
RI PT
ROIs in several SN regions, including between the dACC and the left cerebellum, between the right AI and the left precuneus and left SMG, and between the left PI and the left and right cerebellum. We will now discuss the implications of each of these findings in turn.
The finding that greater FC within the DMN is associated with greater EA is consistent with previous
SC
studies linking activation in DMN regions to both LEAS scores (Frewen et al., 2008; Lane et al., 1998a)
and to emotion-focused attention and emotion recognition (Kalisch et al., 2006; Lane et al., 2015; Peelen et
M AN U
al., 2010; Roy et al., 2012; R Smith et al., 2014b; Smith and Lane, 2015). It is also consistent with the theory of Levels of Emotional Awareness (Lane and Schwartz, 1987), which suggests that higher EA is associated with the ability to recognize one’s own emotional responses using more fine-grained, differentiated conceptual categories (e.g., the concepts “sad” and “angry” are more fine-grained than the concept “bad”). This is because interactions between DMN regions – the MPFC, PCC, and hippocampus in
TE D
particular – are thought to subserve conceptualization of one’s current perceptual experience in light of past experiences stored in long-term memory (Barrett and Satpute, 2013; Lindquist and Barrett, 2012), and therefore one would expect more efficient information exchange between DMN regions (as reflected by greater resting FC) to facilitate more adaptive conceptualization of emotional experience. It is also notable
EP
that previous studies have found positive correlations both between LEAS scores and vagally mediated heart rate variability (HRV; Verkuil, Brosschot, Tollenaar, Lane, & Thayer, 2016) and between HRV and
AC C
activity within medial prefrontal DMN regions (Thayer et al., 2012). Because HRV also represents an important peripheral physiological predictor of both physical and emotional health (La Rovere et al., 1998; Thayer and Lane, 2009), when combined with the results of this study it appears plausible to suggest that the same DMN-mediated conceptualization processes facilitating higher EA may also be among those promoting healthier regulation of emotions and associated bodily responses.
Our findings showing greater resting FC in the SN with greater LEAS scores also support the suggestions offered above. This is because bodily perception/regulation figure centrally in many current neural models
ACCEPTED MANUSCRIPT 12 of emotion (Barrett and Simmons, 2015; Smith and Lane, 2015; Zaki et al., 2012). Specifically, when one considers that the EA-related ability to recognize one’s own emotions may in part depend on recognizing the meaning of one’s own bodily responses (Smith and Lane, 2015), it appears highly plausible that greater
RI PT
connectivity between SN regions involved in interoception should facilitate better EA. For example, if interoceptive processing were hindered by less efficient information exchange, one might expect less
clearly differentiated bodily percepts, which would in turn promote lower EA (e.g., one might be more
likely to perceive bodily feelings as simply being “unpleasant”). It is notable that higher EA was largely
SC
associated with greater FC between cortical (e.g., insula, dACC) and cerebellar regions of the SN. Current models of cortico-cerebellar interactions suggest that, similar to its role in motor control, the cerebellum
M AN U
may function to optimize cortical processing by implementing a predictive internal model of cortical activity (Buckner, 2013). Thus, perhaps our results could reflect more accurate/efficient interoceptive processing in those with high EA as a function of better cerebellar optimization. Given that predictive internal models improve as a result of repeated performance and feedback, better cerebellar optimization of interoceptive processing could in turn reflect more frequent use of interoceptive attention in those with high
TE D
EA. Future studies should further investigate these interesting possibilities.
There are a few important limitations of the present study that should be highlighted. First, because this study only included healthy participants, it is not clear whether these results will generalize to clinical populations. Future studies should therefore test whether the same relationships between LEAS and FC are
EP
present in the context of the emotional and physical health issues to which low LEAS scores have been previously associated (as reviewed in the introduction). For example, given the association between higher
AC C
LEAS and more differentiated reporting of somatic sensations (Lane et al., 2011), one might predict lesser connectivity with cerebellar regions in patients with somatic symptom disorders or in depressed or anxious patients with especially pronounced somatic symptoms. Second, although we controlled for age, gender, and IQ in our analyses, it is noteworthy that our sample had significantly higher mean IQ scores than the population average. Thus, these LEAS-FC relationships should also be confirmed in a more diverse sample including a broader and more representative range of IQ scores. A method for oral administration of the LEAS has been validated for use in those with more limited educational attainment that affects the reading and writing abilities needed in completing the standard LEAS (Roberton et al., 2013).
ACCEPTED MANUSCRIPT 13 It is also important to highlight the correlational nature of our results. Thus, while we have suggested that higher FC may index better information exchange between network regions engaged in interoceptive- and conceptualization-related processing, and that this could in turn promote higher EA, this remains just one of
RI PT
many possible causal interpretations of our correlational results. Future studies will be necessary to disentangle the directions of causal influence that may be involved (e.g., by examining changes in FC as a result of EA training). Finally, it should be emphasized that our hypotheses in this study were based on 1) the theoretical/empirical links between LEAS scores, conceptualization of emotions, and interoception
SC
(e.g., reviewed in Lane et al., 2015), and 2) the theoretical/empirical links between the latter two
psychological processes and the DMN and SN, respectively (e.g., reviewed in Barrett and Satpute, 2013;
M AN U
Lindquist and Barrett, 2012). We have therefore interpreted our neuroimaging results to support the link between LEAS scores and those psychological processes. However, because there does not appear to be a 1-to-1 mapping between brain regions and psychological processes (e.g., see Anderson, 2014), other interpretations of our results may also be possible. Future studies will therefore also be necessary to confirm that the associations we have observed between EA, the DMN, and the SN are best explained by
TE D
the conceptualization- and interoception-related processes that we have suggested.
In conclusion, this study found that greater EA, as measured by the LEAS, was associated with greater FC within the DMN and also within the SN. We have suggested this may index improved perceptual processing of bodily feelings, and improved conceptualization of those feelings in the context of an
EP
emotional reaction. Given that LEAS scores have been associated with several clinical phenomena associated with physical and emotional health, future research should examine these LEAS-FC
AC C
relationships in clinical populations. This could perhaps lead to novel, potentially predictive neural correlates of low EA and associated physical/mental health risk.
References
Anderson, M., 2014. After Phrenology: Neural Reuse and the Interactive Brain. MIT Press, Cambridge, MA.
Barchard, K., Bajgar, J., Leaf, D., Lane, R., 2010. Computer scoring of the Levels of Emotional Awareness
ACCEPTED MANUSCRIPT 14 Scale. Behav. Res. Methods 42, 586–595.
Barchard, K., Hakstian, A., 2004. The nature and measurement of emotional intelligence abilities; basic
Psychol. Meas. 64, 437–462.
RI PT
dimensions and their relationships with other cognitive abilities and personality variables. Educ.
Barrett, L., Lane, R., Sechrest, L., Schwartz, G., 2000. Sex Differences in Emotional Awareness. Personal. Soc. Psychol. Bull. 26, 1027–1035. doi:10.1177/01461672002611001
SC
Barrett, L., Satpute, A., 2013. Large-scale brain networks in affective and social neuroscience: towards an
doi:10.1016/j.conb.2012.12.012
M AN U
integrative functional architecture of the brain. Curr. Opin. Neurobiol. 23, 361–72.
Barrett, L., Simmons, W., 2015. Interoceptive predictions in the brain. Nat. Rev. Neurosci. 16, 419–29. doi:10.1038/nrn3950
Baslet, G., Termini, L., Herbener, E., 2009. Deficits in emotional awareness in schizophrenia and their
TE D
relationship with other measures of functioning. J. Nerv. Ment. Dis. 197, 655.
Behzadi, Y., Restom, K., Liau, J., Liu, T., 2007. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 37, 90–101.
EP
doi:10.1016/j.neuroimage.2007.04.042
Berthoz, S., Ouhayoun, B., Parage, N., 2000. Etude preliminaire des niveaux de conscience emotionnelle
AC C
chez des patients deprimes et des controles. (Preliminary study of the levels of emotional awareness in depressed patients and controls.). Ann. Med. Psychol. 158, 665–672.
Bréjard, V., Bonnet, A., Pedinielli, J., 2012. The role of temperament and emotional awareness in risk taking in adolescents. L’Encéphale Rev. Psychiatr. Clin. Biol. Thérapeutique 38, 1–9.
Buckner, R., 2013. The Cerebellum and Cognitive Function: 25 Years of Insight from Anatomy and Neuroimaging. Neuron 80, 807–815. doi:10.1016/j.neuron.2013.10.044
ACCEPTED MANUSCRIPT 15 Buckner, R., Andrews-Hanna, J.R., Schacter, D.L., 2008. The Brain’s Default Network. Ann. N. Y. Acad. Sci. 1124, 1–38. doi:10.1196/annals.1440.011
Bydlowski, S., Corcos, M., Jeammet, P., Paterniti, S., Berthoz, S., Laurier, C., Chambry, J., Consoli, S.,
RI PT
2005. Emotion-processing deficits in eating disorders. Int. J. Eat. Disord. 37, 321–329.
Chai, X., Nieto-Castanon, A., Öngür, D., Whitfield-Gabrieli, S., 2012. Anticorrelations in resting state networks without global signal regression. Neuroimage 59, 1420–1428.
SC
doi:10.1016/j.neuroimage.2011.08.048
Ciarrochi, J., Caputi, P., Mayer, J., 2003. The distinctiveness and utility of a measure of trait emotional
M AN U
awareness. Pers. Individ. Dif. 34, 1477–1490.
Consoli, S., Lemogne, C., Roch, B., Laurent, S., Plouin, P., Lane, R., 2010. Differences in emotion processing in patients with essential and secondary hypertension. Am. J. Hypertens. 23, 515–521.
Craig, A.D., 2009. How do you feel--now? The anterior insula and human awareness. Nat. Rev. Neurosci.
TE D
10, 59–70.
Critchley, H., 2005. Neural mechanisms of autonomic, affective, and cognitive integration. J. Comp.
EP
Neurol. 493, 154–66. doi:10.1002/cne.20749
Critchley, H., Mathias, C., Josephs, O., O’Doherty, J., Zanini, S., Dewar, B.-K., Cipolotti, L., Shallice, T., Dolan, R., 2003. Human cingulate cortex and autonomic control: converging neuroimaging and
AC C
clinical evidence. Brain 126, 2139–52. doi:10.1093/brain/awg216
Donges, U., Kersting, A., Dannlowski, U., Lalee-Mentzel, J., Arolt, V., Suslow, T., 2005. Reduced awareness of others’ emotions in unipolar depressed patients. J. Nerv. Ment. Dis. 193, 331–337.
Farb, N., Segal, Z., Anderson, A., 2013. Attentional modulation of primary interoceptive and exteroceptive cortices. Cereb. cortex 23, 114–26. doi:10.1093/cercor/bhr385
Frewen, P., Lane, R., Neufeld, R., Densmore, M., Stevens, T., Lanius, R., 2008. Neural correlates of levels
ACCEPTED MANUSCRIPT 16 of emotional awareness during trauma script-imagery in posttraumatic stress disorder. Psychosom. Med. 70, 27–31.
Gusnard, D., Akbudak, E., Shulman, G., Raichle, M., 2001. Medial prefrontal cortex and self-referential
RI PT
mental activity: relation to a default mode of brain function. Proc. Natl. Acad. Sci. 98, 4259–4264.
Kalisch, R., Wiech, K., Critchley, H.D., Dolan, R.J., 2006. Levels of appraisal: a medial prefrontal role in high-level appraisal of emotional material. Neuroimage 30, 1458–66.
SC
doi:10.1016/j.neuroimage.2005.11.011
La Rovere, M., Bigger, J., Marcus, F., Mortara, A., Schwartz, P., 1998. Baroreflex sensitivity and heart-rate
doi:10.1016/S0140-6736(97)11144-8
M AN U
variability in prediction of total cardiac mortality after myocardial infarction. Lancet 351, 478–484.
Lackner, J., 2005. Is IBS a problem of emotion dysregulation? Testing the levels of emotional awareness model, in: Presented at the Annual Meeting of the American Psychosomatic Society.
TE D
Lane, R., Carmichael, C., Reis, H., 2011. Differentiation in the momentary rating of somatic symptoms covaries with trait emotional awareness in patients at risk for sudden cardiac death. Psychosom. Med. 73, 185–92. doi:10.1097/PSY.0b013e318203b86a
EP
Lane, R., Quinlan, D., Schwartz, G., Walker, P., Zeitlin, S., 1990. The Levels of Emotional Awareness Scale: a cognitive-developmental measure of emotion. J. Pers. Assess. 55, 124–34.
AC C
doi:10.1080/00223891.1990.9674052
Lane, R., Reiman, E., Axelrod, B., Yun, L., Holmes, A., Schwartz, G., 1998a. Neural correlates of levels of emotional awareness. Evidence of an interaction between emotion and attention in the anterior cingulate cortex. J. Cogn. Neurosci. 10, 525–535.
Lane, R., Schwartz, G., 1987. Levels of emotional awareness: a cognitive-developmental theory and its application to psychopathology. Am. J. Psychiatry 144, 133–143.
ACCEPTED MANUSCRIPT 17 Lane, R., Sechrest, L., Reidel, R., Weldon, V., Kaszniak, A., Schwartz, G., 1996. Impaired verbal and nonverbal emotion recognition in alexithymia. Psychosom. Med. 58, 203–10.
Lane, R., Sechrest, L., Riedel, R., 1998b. Sociodemographic correlates of alexithymia. Compr. Psychiatry
RI PT
39, 377–385. doi:10.1016/S0010-440X(98)90051-7
Lane, R., Sechrest, L., Riedel, R., Shapiro, D., Kaszniak, A., 2000. Pervasive emotion recognition deficit common to alexithymia and the repressive coping style. Psychosom. Med. 62, 492–501.
SC
Lane, R., Weihs, K., Herring, A., Hishaw, A., Smith, R., 2015. Affective agnosia: Expansion of the alexithymia construct and a new opportunity to integrate and extend Freud’s legacy. Neurosci.
M AN U
Biobehav. Rev. 55, In Press. doi:10.1016/j.neubiorev.2015.06.007
Levine, D., Marziali, E., Hood, J., 1997. Emotion processing in borderline personality disorders. J. Nerv. Ment. Dis. 185, 240–246.
Li, W., Mai, X., Liu, C., 2014. The default mode network and social understanding of others: what do brain
TE D
connectivity studies tell us. Front. Hum. Neurosci. 8, 74. doi:10.3389/fnhum.2014.00074
Lindquist, K., Barrett, L., 2012. A functional architecture of the human brain: emerging insights from the
EP
science of emotion. Trends Cogn. Sci. 16, 533–540. doi:10.1016/j.tics.2012.09.005
McRae, K., Reiman, E., Fort, C., Chen, K., Lane, R., 2008. Association between trait emotional awareness and dorsal anterior cingulate activity during emotion is arousal-dependent. Neuroimage 41, 648–55.
AC C
doi:10.1016/j.neuroimage.2008.02.030
Medford, N., Critchley, H., 2010. Conjoint activity of anterior insular and anterior cingulate cortex: awareness and response. Brain Struct. Funct. 214, 535–549. doi:10.1007/s00429-010-0265-x
Moeller, S., Konova, A., Parvaz, M., Tomasi, D., Lane, R., Fort, C., Goldstein, R., RZ, G., 2014. Functional, Structural, and Emotional Correlates of Impaired Insight in Cocaine Addiction. JAMA Psychiatry 71, 61. doi:10.1001/jamapsychiatry.2013.2833
ACCEPTED MANUSCRIPT 18 Peelen, M., Atkinson, A., Vuilleumier, P., 2010. Supramodal representations of perceived emotions in the human brain. J. Neurosci. 51, 10127–34. doi:10.1523/JNEUROSCI.2161-10.2010
Raichle, M., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001. A default
RI PT
mode of brain function. Proc Natl Acad Sci U S A 98, 676–682. doi:10.1073/pnas.98.2.676
Roberton, T., Daffern, M., Bucks, R., 2013. Oral administration of the Levels of Emotional Awareness Scale. Aust. J. Psychol. 65, 172–179. doi:10.1111/ajpy.12018
SC
Roy, M., Shohamy, D., Wager, T.D., 2012. Ventromedial prefrontal-subcortical systems and the generation
10.1016/j.tics.2012.01.005
M AN U
of affective meaning. Trends Cogn. Sci. 16, 147–156. doi:S1364-6613(12)00027-7 [pii]
Shirer, W., Ryali, S., Rykhlevskaia, E., Menon, V., Greicius, M., 2012. Decoding subject-driven cognitive states with whole-brain connectivity patterns. Cereb. cortex 22, 158–65. doi:10.1093/cercor/bhr099
Smith, R., Braden, B., Chen, K., Ponce, F., Lane, R., Baxter, L., 2014a. The neural basis of attaining
TE D
conscious awareness of sad mood. Brain Imaging Behav. 9. doi:10.1007/s11682-014-9318-8
Smith, R., Fass, H., Lane, R., 2014b. Role of medial prefrontal cortex in representing one’s own subjective emotional responses: A preliminary study. Conscious. Cogn. 29, 117–130.
EP
doi:10.1016/j.concog.2014.08.002
Smith, R., Lane, R., 2015. The neural basis of one’s own conscious and unconscious emotional states.
AC C
Neurosci. Biobehav. Rev. 57, 1–29. doi:10.1016/j.neubiorev.2015.08.003
Subic-Wrana, A., Beetz, M., Paulussen, J., Wiltnik, J., Beutel, M., 2007. Relations between attachment, childhood trauma, and emotional awareness in psychosomatic inpatients, in: Presented at the Annual Meeting of the American Psychosomatic Society.
Subic-Wrana, C., Bruder, S., Thomas, W., Lane, R., Köhle, K., 2005. Emotional awareness deficits in inpatients of a psychosomatic ward: a comparison of two different measures of alexithymia.
ACCEPTED MANUSCRIPT 19 Psychosom. Med. 67, 483–489.
Tavares, P., Barnard, P., Lawrence, A., 2011. Emotional complexity and the neural representation of
RI PT
emotion in motion. Soc. Cogn. Affect. Neurosci. 6, 98–108. doi:10.1093/scan/nsq021
Thayer, J., Ahs, F., Fredrikson, M., Sollers, J.J., Wager, T.D., 2012. A meta-analysis of heart rate
variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health. Neurosci. Biobehav. Rev. 36, 747–56.
SC
Thayer, J., Lane, R., 2009. Claude Bernard and the heart-brain connection: further elaboration of a model
M AN U
of neurovisceral integration. Neurosci. Biobehav. Rev. 33, 81–88.
Verkuil, B., Brosschot, J., Tollenaar, M., Lane, R., Thayer, J., 2016. Prolonged Non-metabolic Heart Rate Variability Reduction as a Physiological Marker of Psychological Stress in Daily Life. Ann. Behav. Med. 1–11. doi:10.1007/s12160-016-9795-7
TE D
Wechsler, D., 2011. WASI -II: Wechsler abbreviated scale of intelligence - second edition, WASI.
Whitfield-Gabrieli, S., Nieto-Castanon, A., 2012. Conn: A functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect. 0, 125–41. doi:10.1089/brain.2012.0073
EP
Whitfield-Gabrieli, S., Thermenos, H., Milanovic, S., Tsuang, M., Faraone, S., McCarley, R., Shenton, M., Green, A., Nieto-Castanon, A., LaViolette, P., Wojcik, J., Gabrieli, J., Seidman, L., 2009. Hyperactivity and hyperconnectivity of the default network in schizophrenia and in first-degree
AC C
relatives of persons with schizophrenia 106, 1279–1284. doi:10.1073/pnas.0809141106
Zaki, J., Davis, J.I., Ochsner, K., 2012. Overlapping activity in anterior insula during interoception and emotional experience. Neuroimage 62, 493–499. doi:10.1016/j.neuroimage.2012.05.012
ACCEPTED MANUSCRIPT 20 Figure Legends
Figure 1. Visual depiction of some of the regions of interest (ROIs) within the DMN (Top; Red) and SN (Bottom; Anterior Network = Yellow, Posterior Network = Blue) that were used in the analyses described
RI PT
in the text. DMN ROIs: MPFC = medial prefrontal cortex (also covers portions of the anterior cingulate
cortex and orbitofrontal cortex), PCC = posterior cingulate cortex (also covers portions of the precuneus), MCC = mid-cingulate cortex, thal = thalamus, L Hc = left hippocampus, R Hc = right hippocampus. SN
ROIs: dACC = dorsal anterior cingulate cortex (also covers portions of dorsal MPFC and supplementary
SC
motor area), R AI = right anterior insula, L AI = left anterior insula, R PI = right posterior insula, L PI =
M AN U
left posterior insula.
Figure 2. Visual depiction of the regions within the DMN (left) and SN (right) that showed a significant positive correlation between functional connectivity and LEAS scores (after statistically accounting for age, gender, and IQ score). Seed regions are labeled in black; target regions are labeled in red. DMN ROIs: MPFC = medial prefrontal cortex (also covers portions of the anterior cingulate cortex and orbitofrontal cortex), PCC = posterior cingulate cortex (also covers portions of the precuneus), MCC = mid-cingulate
TE D
cortex, thal = thalamus, L AG = left angular gyrus, L Hc = left hippocampus, R Hc = right hippocampus. SN ROIs: dACC = dorsal anterior cingulate cortex (also covers portions of dorsal MPFC and supplementary motor area), R AI = right anterior insula, L PI = left posterior insula, L PreC = left
EP
precuneus, L SMG = left supramarginal gyrus (also covers portions of the inferior parietal gyrus), L cer (A) = left cerebellum (anterior network ROI), L cer (P) = left cerebellum (posterior network ROI), R cer (A) =
AC C
right cerebellum (anterior network ROI).
Figure 3. Example scatter plots illustrating the positive relationship between LEAS total scores and functional connectivity values within four ROI seed-target pairs (i.e., two from each network analysis). See Tables 1 and 2 for associated statistics (i.e., which also account for included covariates). MPFC = medial prefrontal cortex; PCC = posterior cingulate cortex; dACC = dorsal anterior cingulate cortex; R AI = right anterior insula.
ACCEPTED MANUSCRIPT
Table 1. Functional Connectivity Results within the Default Mode Network: Positive Correlations with LEAS Scores 1
RI PT
Table 1.1: Seed ROI = MPFC
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
PCC
2.90
p = .017
.82 (+/- .18)
Right Hippocampus
2.62
p = .025
M AN U
SC
Target ROI
Tscore
.26 (+/- .14)
One-Sample Ttest*
t = 23.14, p < .001
t = 9.62, p < .001
TE D
Left Angular Gyrus 2.38 p = .035 .57 (+/- .25) t = 11.62, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0). Table 1.2: Seed ROI = PCC
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
EP
Target ROI
Tscore
3.68
p = .011
.28 (+/- .14)
t = 10.51, p < .001
Bilateral Thalamus
2.95
p = .017
.04 (+/- .18)
t = 1.17, p = .26
AC C
Left Hippocampus
1Peak
Threshold = P < 0.001, Uncorrected; Cluster Threshold = P< 0.05, FDRCorrected
ACCEPTED MANUSCRIPT
2.90
p = .017
.82 (+/- .18)
Mid-Cingulate
2.20
p = .043
.46 (+/- .17)
t = 23.14, p < .001
RI PT
MPFC
t = 13.42, p < .001
SC
Right Hippocampus 2.15 p = .043 .23 (+/- .13) t = 8.97, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
Table 2.1: Seed ROI = dACC
M AN U
Table 2. Functional Connectivity Results within the Salience Network: Positive Correlations with LEAS Scores 2
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
Left Cerebellum (Anterior Salience Network)
3.82
p = .011
.32 (+/- .14)
t = 11.30, p < .001
TE D
Target ROI
Tscore
AC C
EP
Left Cerebellum (Posterior Salience Network) 3.80 p = .011 .10 (+/- .16) t = 3.06, p = .005 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0). Table 2.2: Seed ROI = right AI
Target ROI
2Peak
Tscore
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
Threshold = P < 0.001, Uncorrected; Cluster Threshold = P< 0.05, FDRCorrected
ACCEPTED MANUSCRIPT
3.12
p = .039
.19 (+/- .23)
t = 4.16, p < .001
RI PT
Left Precuneus
Left SMG 3.11 p = .039 .44 (+/- .25) t = 9.17, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
Tscore
Left Cerebellum (Anterior Salience Network)
4.91
p = .003
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
.08 (+/- .13)
t = 3.04, p = .006
M AN U
Target ROI
p-value (FDRCorrected)
SC
Table 2.3: Seed ROI = left PI
AC C
EP
TE D
Right Cerebellum (Anterior Salience Network) 3.80 p = .011 .00 (+/- .12) t = .004, p = .996 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S1053-8119(17)30612-2
DOI:
10.1016/j.neuroimage.2017.07.044
Reference:
YNIMG 14209
To appear in:
NeuroImage
Received Date: 24 January 2017 Revised Date:
16 June 2017
Accepted Date: 19 July 2017
Please cite this article as: Smith, R., Alkozei, A., Bao, J., Smith, C., Lane, R.D., Killgore, W.D.S., Resting state functional connectivity correlates of emotional awareness, NeuroImage (2017), doi: 10.1016/ j.neuroimage.2017.07.044. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Resting State Functional Connectivity Correlates of Emotional Awareness Ryan Smith1, Anna Alkozei1, Jennifer Bao1, Courtney Smith1, Richard D. Lane1, William D. S. Killgore1
Department of Psychiatry, University of Arizona, 1501 N. Campbell Ave, Tucson, AZ 85724-5002,
RI PT
1
Corresponding author at:
TE D
M AN U
SC
United States
EP
University of Arizona, Department of Psychiatry, 1501 N., Campbell Ave., PO Box 245002, Room 7304B, Tucson, AZ 85724-5002, United States. Tel.: +1 602 501 4168; fax: +1 520 626 6050.
AC C
E-mail address: [email protected] (Ryan Smith).
ACCEPTED MANUSCRIPT 2 Abstract Multiple neuroimaging studies have now linked emotional awareness (EA), as measured by the Levels of Emotional Awareness Scale (LEAS), with activation in regions of neural networks associated with both
RI PT
conceptualization (i.e., default mode network [DMN] regions) and interoception (i.e., salience network [SN] regions) – consistent with the definition of EA as one’s ability to appropriately recognize,
conceptualize, and articulate the emotions of self and other in fine-grained, differentiated ways. However, no study has yet tested the hypothesis that greater LEAS scores are associated with greater resting state
SC
functional connectivity (FC) within these networks. Twenty-six adults (13 female) underwent resting state functional magnetic resonance imaging, and also completed the LEAS. Using pre-defined functional ROIs
M AN U
from the DMN and SN, we observed that LEAS scores were significantly positively correlated with FC between several regions of both of these networks, even when controlling for differences in general intelligence (IQ). These results suggest that higher EA may be associated with more efficient information exchange between brain regions involved in both interoception- and conceptualization-based processing, which could plausibly contribute to more differentiated bodily feelings and more fine-grained
TE D
conceptualization of those feelings.
Keywords: Emotional Awareness; Levels of Emotional Awareness Scale (LEAS); Default Mode Network;
AC C
EP
Salience Network; Conceptualization; Interoception
ACCEPTED MANUSCRIPT 3 Introduction
It is a commonplace clinical observation that there are substantial individual differences in people’s awareness of their own emotions. Specifically, some individuals appear to have a greater ability than others
RI PT
to understand, and clearly articulate, the emotions they are feeling. One well-established and validated
measure of these individual differences in emotional awareness (EA) is the Levels of Emotional Awareness Scale (LEAS) (Lane et al., 1990; Lane and Schwartz, 1987). As measured by the LEAS, higher EA has been found to correlate positively with many adaptive traits/abilities, including self-reported impulse
SC
control (Bréjard et al., 2012), openness to feelings (Lane et al., 1990), emotion recognition ability (Lane et al., 2000, 1996), empathy (Barchard and Hakstian, 2004), and a stable sense of well-being independent of
M AN U
current mood (Ciarrochi et al., 2003). In contrast, lower LEAS scores have been associated with several maladaptive clinical phenomena, such as essential hypertension (Consoli et al., 2010), eating disorders (Bydlowski et al., 2005), post-traumatic stress disorder (Frewen et al., 2008), schizophrenia (Baslet et al., 2009), depression (Berthoz et al., 2000; Donges et al., 2005), borderline personality disorder (Levine et al., 1997), somatoform disorders (Subic-Wrana et al., 2005), a “disorganized attachment style” (Subic-Wrana
TE D
et al., 2007), impaired insight in the context of cocaine abuse (Moeller et al., 2014), and greater pain in patients with irritable bowel syndrome (IBS) (Lackner, 2005). Thus, EA appears to represent an important individual difference variable associated with both emotional and physical health.
EP
The neural basis of EA has also begun to receive experimental investigation. For example, two different task-based functional neuroimaging studies have demonstrated that higher LEAS scores are associated with
AC C
greater activity in the dorsal anterior cingulate cortex (dACC) (Lane et al., 1998a; McRae et al., 2008) – a region of the “salience network” (SN; Barrett & Satpute, 2013) implicated in autonomic regulation (Critchley et al., 2003), attention to one’s bodily states (Farb et al., 2013), and facilitating the influence of bodily feelings on action selection (Critchley, 2005; Medford and Critchley, 2010). Another study has also shown that, during recall of life-threatening experiences, healthy subjects showed greater activity in the rostral ACC (rACC) as a function of greater scores on the LEAS (Frewen et al., 2008). The rACC, and adjacent parts of the medial prefrontal cortex (MPFC), are known to comprise a significant hub of the “default mode network” (DMN) (Barrett and Satpute, 2013; Buckner et al., 2008; Li et al., 2014; Raichle et
ACCEPTED MANUSCRIPT 4 al., 2001), which refers to a set of brain regions whose activity is highly correlated at rest and which are thought to play an important role in the process of conceptualization. In combination with the results of other studies of the rACC/MPFC (Kalisch et al., 2006; Peelen et al., 2010; Roy et al., 2012; R Smith et al.,
RI PT
2014b), we have previously suggested that this region, in conjunction with the rest of the DMN, may play an important role in assigning conceptual significance to bodily feelings (Lane et al., 2015; Smith and
Lane, 2015) – and therefore allow such feelings to be understood in explicit emotional terms (e.g., sadness, fear, etc.). This suggestion is also consistent with another study (Tavares et al., 2011), which showed that,
SC
while subjects viewed simple animated scenarios with social/emotional content, higher LEAS scores
predicted greater neural activity within abstract semantic processing regions (i.e., left anterior temporal
M AN U
cortex), whereas lower LEAS scores predicted more concrete action-oriented brain activation (i.e., in premotor cortex).
While the studies described above have examined the association between EA and task-related neural responses, it is notable that no study to date has yet examined the relation between LEAS scores and resting state measures of functional connectivity (FC) in the DMN or SN. Given that DMN regions have
TE D
themselves been largely characterized using resting state connectivity measures (Buckner et al., 2008), and given the theoretical/empirical work described above supporting a link between EA and DMN regions/functions, there are strong reasons to hypothesize that LEAS scores should predict individual differences in DMN connectivity at rest. Given that EA also involves recognition/articulation of
EP
differentiated bodily feeling states, and that the SN is involved in representing bodily percepts (and using them to guide cognition and action selection), there is also reason to predict LEAS scores would be
AC C
associated with better FC between SN regions. In the present study, we therefore examined the correlations between LEAS scores and individual differences in resting state FC with the hypothesis that higher LEAS scores would be associated with stronger positive connectivity between DMN regions and SN regions, respectively.
Materials and Methods
Participants
ACCEPTED MANUSCRIPT 5 Twenty-six adults (13 female; mean age = 23.12 +/- 4.03) were recruited from the general population via flyers and internet advertisements to participate in the present study. Participants did not have any history of psychiatric or neurological disorders (assessed via a phone screen questionnaire based on criteria within
RI PT
the Diagnostic and Statistical Manual for Mental Disorders, 4th addition; DSM-IV-TR), and all provided written informed consent prior to participation. All participants received a nominal financial compensation for participation. The research protocol of the present study was also reviewed and approved by the Institutional Review Board of the University of Arizona.
SC
Procedure
M AN U
Upon completing the informed consent process, participants were taken to the magnetic resonance imaging (MRI) scanner at the University of Arizona where they underwent a resting state functional scan (see Neuroimaging Methods below). After completing the resting state functional scan, participants were escorted back to the lab, seated at a laptop, and asked to complete an on-line version of the LEAS (www.eleastest.net) that makes use of a validated automatic scoring program (Barchard et al., 2010).
TE D
LEAS. The LEAS presents participants with 2-4 sentence descriptions of 20 social situations, where each situation includes 2 people. The situation descriptions are designed to elicit four types of emotion (sadness, happiness, anger, and fear) at 5 levels of complexity. One situation is presented on each electronically presented page, followed by two questions: “How would you feel?” and “How would the other person
EP
feel?” Separate response boxes are provided for typing in the answers to each question. Participants are instructed to type their responses into these boxes, and they are asked to use as much or as little space as
AC C
needed to answer. The only rule given is that they must use the word “feel” in their responses.
Scores reflecting EA level are assigned based on the words participants provide in their responses. The lowest scores (Level 0) are given to words that do not refer to feelings. Level 1 scores are given to words that refer to physiological sensations (e.g., “tired”), whereas level 2 scores are given to words referring to feeling-related actions (e.g., “punching”) or simple valence discriminations (e.g., “bad,” “good”) that have inherent avoidance- or approach-related content. Level 3 scores are assigned to words referring to single emotion concepts (e.g., “happy,” “sad”). Level 4 scores are given when at least 2 words from level 3 are
ACCEPTED MANUSCRIPT 6 used (i.e., when they convey greater emotional differentiation than either word alone). For each item, the self- and other-related responses are scored separately (i.e., with a value of 0-4). A “total” score is also given for each of the 20 LEAS items; this score reflects the higher of the self- and other-related scores,
RI PT
unless a score of 4 is given for both. In this case, a total score of 5 is given for the item, so long as the selfand other-related responses are sufficiently differentiable (for more detail, see Lane et al., 1990).
General Intelligence. Intelligence quotient (IQ) was measured with the two-subtest form (FSIQ-2) of the Wechsler Abbreviated Scale of Intelligence – Second Edition (WASI-II; Pearson Assessment, Inc., San
SC
Antonio, TX; Wechsler, 2011). This was done in order to control for general intelligence when examining
M AN U
FC correlates of LEAS scores.
Neuroimaging Methods
Neuroimaging was performed within a 3T Siemens Skyra scanner (Siemens, Erlangen, Germany) with a 32-channel head coil. T1-weighted structural images (3D MPRAGE) were acquired (TR/TE/flip angle = 2.1 s / 2.33 ms/ 12 degree) covering 176 sagittal slices (256 x 256) with a slice thickness of 1 mm (voxel
TE D
size = 1 x 1 x 1). Functional T2*-weighted scans were acquired over 32 transverse slices (2.5 mm thickness; matrix: 88x84). Each volume was collected with an interleaved sequence (TR/TE/flip angle = 2 s/ 25 ms/ 90 degree). The voxel size of the T2* sequence was 2.5 x 2.5 x 3.5 mm (i.e., with a 40% slice
240 mm.
EP
gap, allowing collection of 300 volumes within a 10-minute acquisition time). The field of view (FOV) was
AC C
Resting-state preprocessing
The publicly available CONN functional connectivity toolbox (version 16.a; https://www.nitrc.org/projects/conn), in conjunction with SPM12 (Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion.ucl.ac.uk/spm), was used to perform all preprocessing steps (using CONN’s default preprocessing pipeline), as well as subsequent statistical analyses, on all collected MRI scans. In this preprocessing pipeline, raw functional images were slice-time corrected, realigned (motion corrected), unwarped, and coregistered to each subject’s MPRAGE image in accordance with standard algorithms. Images were then normalized to Montreal Neurological Institute (MNI) coordinate
ACCEPTED MANUSCRIPT 7 space, spatially smoothed (8 mm full-width at half maximum), and resliced to 2 x 2 x 2 mm voxels. The Artifact Detection Tool (ART; http://www.nitrc.org/projects/artifact_detect/) was also used to regress out scans as nuisance covariates in the first-level analysis exceeding 3 SD in mean global intensity and scan-to-
RI PT
scan motion that exceeded 0.5 mm. These were added in addition to covariates for the 6 rotation/translation movement parameters.
Functional Connectivity Analysis
SC
Using a standard seed-driven approach, FC analyses were performed using the default FC processing pipeline in the CONN toolbox (for details, see Whitfield-Gabrieli & Nieto-Castanon, 2012). In this
M AN U
processing pipeline, physiological and other spurious sources of noise were estimated with the aCompcor method (Behzadi et al., 2007; Chai et al., 2012; Whitfield-Gabrieli et al., 2009); they were then removed together with the movement- and artifact-related covariates mentioned above. The residual BOLD timeseries was then band-pass filtered (.008Hz–.09 Hz). Every participant's structural image was segmented into gray matter, white matter, and cerebral spinal fluid using SPM12. White matter and cerebral spinal fluid noise ROIs were removed through regression. The ROIs we used were taken from a freely available
TE D
atlas of regions defined by correlated activation patterns (http://findlab.stanford.edu/functional_ROIs.html). This atlas includes 90 ROIs, grouped into several networks: The anterior and posterior SN, the dorsal and ventral DMN, the left and right executive control network, as well as the auditory network, basal ganglia
EP
network, higher visual network, language network, sensorimotor network, primary visual network, visuospatial network, and precuneus network (for details, see Shirer, Ryali, Rykhlevskaia, Menon, &
AC C
Greicius, 2012). However, we only used the dorsal DMN and anterior/posterior SN ROIs in our analyses.
To produce first-level correlation maps, for each participant the residual BOLD time course was then extracted from each DMN and SN ROI and Pearson's correlation coefficients were computed between a priori selected seed ROI time courses (i.e., 2 seed ROIs from the DMN and 5 seed ROIs from the SN; see the paragraph immediately below for further explanation) and the time courses of all other included ROIs. The resulting correlation coefficients (i.e., one for each seed-target ROI pair in each participant) were then Fisher transformed into ‘Z’ scores to increase normality and thus improve the subsequent second-level General Linear Model analyses.
ACCEPTED MANUSCRIPT 8 At the second level, we then performed two analyses based on our a priori hypotheses (i.e., by compiling/examining the Fisher-transformed Z scores from each participant described above). The first analysis tested the hypothesis that the ROIs within the dorsal DMN would show higher connectivity in
RI PT
individuals with higher LEAS scores. The dorsal DMN includes 9 ROIs (see Figure 1): the MPFC (this ROI also covers parts of the ACC and orbitofrontal cortex [OFC]), the posterior cingulate (PCC; this ROI also covers parts of the precuneus), left hippocampus, a bilateral region of the thalamus, right hippocampus, a region of the right superior frontal gyrus, the mid-cingulate, left angular gyrus, and the right angular
SC
gyrus. For this analysis we chose to use the MPFC ROI and the PCC ROI as seeds, as these are recognized as central hubs within this network and have also been most strongly linked to emotional
M AN U
attention/awareness in previous studies (e.g., Frewen et al., 2008; Gusnard et al., 2001; Lane et al., 2015; R Smith et al., 2014a, 2014b; Smith and Lane, 2015). Connectivity was tested between these two seed ROIs and all other dorsal DMN ROIs listed above. The second analysis tested the hypothesis that the ROIs within the SN (both anterior and posterior network regions together) would have higher connectivity in individuals with higher LEAS scores. The anterior portion of the SN included 7 ROIs (see Figure 1): the left and right anterior insula (AI), the dACC (this ROI also covers parts of the dorsal MPFC and
TE D
supplementary motor area; SMA), regions of the left and right middle frontal gyrus, and regions of the left and right cerebellum. The posterior portion of the SN included 12 ROIs (see Figure 1), including the left and right posterior insula (PI), the left and right supramarginal gyrus (SMG; this ROI also covers parts of
EP
the inferior parietal gyrus [IPG]), regions of the left and right precuneus, regions of the left and right cerebellum, regions of the left and right thalamus, a region of the left middle frontal gyrus, and a region
AC C
within the right mid-cingulate (Shirer, Ryali, Rykhlevskaia, Menon, & Greicius, 2012). As seeds in this analysis, we used the left and right AI, left and right PI, and the dACC ROIs, as the insula and dACC have been most strongly linked to emotional awareness/attention in previous studies (e.g., Craig, 2009; Lane et al., 1998a; McRae et al., 2008; Medford and Critchley, 2010; R Smith et al., 2014a; Smith and Lane, 2015). Connectivity was tested between these five seed ROIs and all other SN ROIs. For these analyses we used an FDR-corrected threshold of p < .05. This correction was applied at the analysis-level, after testing for connectivity between all seed-target ROI pairs in a given network. In these analyses we also included age, gender, and FSIQ-2 scores as covariates of no interest. This was done to ensure that our results only
ACCEPTED MANUSCRIPT 9 revealed FC differences that were explained uniquely by EA, and not attributable to differences in general intelligence, age, or gender (e.g., as suggested by previous work; see Barrett, Lane, Sechrest, & Schwartz, 2000; Lane, Quinlan, Schwartz, Walker, & Zeitlin, 1990b; Lane, Sechrest, & Riedel, 1998; Roberton,
RI PT
Daffern, & Bucks, 2013).
In a final exploratory analysis, we also examined connectivity between each of the DMN and SN hub
regions used as seed regions in the above analyses. This between-network analysis was done to explore
whether differences in EA might also be associated with differences in how the DMN and SN functionally
SC
interact. This analysis used the same covariates and statistical thresholds stated above, and examined
whether LEAS scores were significantly associated with connectivity values between the following ROIs
M AN U
(i.e., all used as both seeds and targets): MPFC, PCC, dACC, left AI, right AI, left PI, and right PI.
Results
Behavioral Results
LEAS scores were as follows: TOTAL = 74.0 (+9.8)%, SELF = 63.0 (+8.7)%, OTHER = 58.5 (+10.7)%.
TE D
Females had numerically higher LEAS TOTAL scores than males (means: females = 76.8 + 10.6; males = 71.2 + 8.4), but this difference was non-significant (t = 1.5, p = .147). The correlation between age and
EP
LEAS scores was negative and non-significant (r = -.315, p = .12).
WASI-II FSIQ-2 scores had a mean of 115.23 (+11.7). LEAS TOTAL scores were significantly positively
AC C
correlated with FSIQ-2 scores (r = .619, p = .001).1
Default Mode Network Analysis
With higher LEAS scores, the MPFC ROI showed greater FC with the PCC, the right hippocampus, and with the left angular gyrus (See Table 1.1; Figures 2 and 3), even after accounting for covariates of age, gender, and FSIQ. With higher LEAS scores, the PCC ROI showed greater FC with the MPFC, the mid-
1 The LEAS scores and FSIQ-2 scores from this data set have previously been published in conjunction with task-related imaging data (Smith et al., in press.). Their relation to resting state imaging data, however, is novel to the present manuscript.
ACCEPTED MANUSCRIPT 10 cingulate, the left and right hippocampus, and with bilateral thalamus (See Table 1.2; Figures 2 and 3). As also reported in Table 1, the mean connectivity values for all but one of these seed-target pairs were significantly greater than zero. The only exception was mean connectivity between the PCC and the
RI PT
bilateral thalamus, which was numerically positive but not significantly different from zero.
Salience Network Analysis
With higher LEAS scores, the dACC ROI showed greater FC with two ROIs in the left cerebellum (i.e.,
SC
one from the anterior and one from the posterior SN; See Table 2.1; Figures 2 and 3). Higher LEAS scores were associated with greater FC between the right AI ROI and the left precuneus and left SMG (i.e., both
M AN U
from the posterior SN; See Table 2.2). With higher LEAS scores, the left PI ROI showed greater FC with the left and right cerebellum ROIs from the anterior SN (See Table 2.3; Figures 2 and 3) 2. The left AI and right PI seeds did not show greater FC with any other SN ROIs with increasing LEAS scores. As also reported in Table 2, the mean connectivity values for all but one of these seed-target pairs were significantly greater than zero. The only exception was mean connectivity between the left PI and the right
Between Network Analysis
TE D
cerebellum, which was not significantly different from zero.
No statistically significant associations were observed between LEAS scores and FC between any of the
EP
seed and target ROIs examined in this analysis (i.e., between the two DMN ROIs and the five SN ROIs used in the other analyses described above).
AC C
Discussion
In this study we tested the hypotheses that greater EA would be associated with stronger resting state FC between regions of pre-established cortical networks subserving 1) conceptualization (i.e., the DMN) and 2) processing of afferent bodily signals (i.e., the SN) (Barrett and Satpute, 2013). In support of the first hypothesis, we found that higher EA (as measured by LEAS scores) predicted stronger FC between ROIs in
2
Tables 1-2 only report significant results. For the full results of each of the analyses, including nonsignificant results, the reader is referred to the supplementary materials.
ACCEPTED MANUSCRIPT 11 several DMN regions, including between the MPFC and the PCC, right hippocampus, and left angular gyrus, as well as between the PCC and the MPFC, the left and right hippocampus, the mid-cingulate, and the thalamus. In support of the second hypothesis, we found that higher EA predicted stronger FC between
RI PT
ROIs in several SN regions, including between the dACC and the left cerebellum, between the right AI and the left precuneus and left SMG, and between the left PI and the left and right cerebellum. We will now discuss the implications of each of these findings in turn.
The finding that greater FC within the DMN is associated with greater EA is consistent with previous
SC
studies linking activation in DMN regions to both LEAS scores (Frewen et al., 2008; Lane et al., 1998a)
and to emotion-focused attention and emotion recognition (Kalisch et al., 2006; Lane et al., 2015; Peelen et
M AN U
al., 2010; Roy et al., 2012; R Smith et al., 2014b; Smith and Lane, 2015). It is also consistent with the theory of Levels of Emotional Awareness (Lane and Schwartz, 1987), which suggests that higher EA is associated with the ability to recognize one’s own emotional responses using more fine-grained, differentiated conceptual categories (e.g., the concepts “sad” and “angry” are more fine-grained than the concept “bad”). This is because interactions between DMN regions – the MPFC, PCC, and hippocampus in
TE D
particular – are thought to subserve conceptualization of one’s current perceptual experience in light of past experiences stored in long-term memory (Barrett and Satpute, 2013; Lindquist and Barrett, 2012), and therefore one would expect more efficient information exchange between DMN regions (as reflected by greater resting FC) to facilitate more adaptive conceptualization of emotional experience. It is also notable
EP
that previous studies have found positive correlations both between LEAS scores and vagally mediated heart rate variability (HRV; Verkuil, Brosschot, Tollenaar, Lane, & Thayer, 2016) and between HRV and
AC C
activity within medial prefrontal DMN regions (Thayer et al., 2012). Because HRV also represents an important peripheral physiological predictor of both physical and emotional health (La Rovere et al., 1998; Thayer and Lane, 2009), when combined with the results of this study it appears plausible to suggest that the same DMN-mediated conceptualization processes facilitating higher EA may also be among those promoting healthier regulation of emotions and associated bodily responses.
Our findings showing greater resting FC in the SN with greater LEAS scores also support the suggestions offered above. This is because bodily perception/regulation figure centrally in many current neural models
ACCEPTED MANUSCRIPT 12 of emotion (Barrett and Simmons, 2015; Smith and Lane, 2015; Zaki et al., 2012). Specifically, when one considers that the EA-related ability to recognize one’s own emotions may in part depend on recognizing the meaning of one’s own bodily responses (Smith and Lane, 2015), it appears highly plausible that greater
RI PT
connectivity between SN regions involved in interoception should facilitate better EA. For example, if interoceptive processing were hindered by less efficient information exchange, one might expect less
clearly differentiated bodily percepts, which would in turn promote lower EA (e.g., one might be more
likely to perceive bodily feelings as simply being “unpleasant”). It is notable that higher EA was largely
SC
associated with greater FC between cortical (e.g., insula, dACC) and cerebellar regions of the SN. Current models of cortico-cerebellar interactions suggest that, similar to its role in motor control, the cerebellum
M AN U
may function to optimize cortical processing by implementing a predictive internal model of cortical activity (Buckner, 2013). Thus, perhaps our results could reflect more accurate/efficient interoceptive processing in those with high EA as a function of better cerebellar optimization. Given that predictive internal models improve as a result of repeated performance and feedback, better cerebellar optimization of interoceptive processing could in turn reflect more frequent use of interoceptive attention in those with high
TE D
EA. Future studies should further investigate these interesting possibilities.
There are a few important limitations of the present study that should be highlighted. First, because this study only included healthy participants, it is not clear whether these results will generalize to clinical populations. Future studies should therefore test whether the same relationships between LEAS and FC are
EP
present in the context of the emotional and physical health issues to which low LEAS scores have been previously associated (as reviewed in the introduction). For example, given the association between higher
AC C
LEAS and more differentiated reporting of somatic sensations (Lane et al., 2011), one might predict lesser connectivity with cerebellar regions in patients with somatic symptom disorders or in depressed or anxious patients with especially pronounced somatic symptoms. Second, although we controlled for age, gender, and IQ in our analyses, it is noteworthy that our sample had significantly higher mean IQ scores than the population average. Thus, these LEAS-FC relationships should also be confirmed in a more diverse sample including a broader and more representative range of IQ scores. A method for oral administration of the LEAS has been validated for use in those with more limited educational attainment that affects the reading and writing abilities needed in completing the standard LEAS (Roberton et al., 2013).
ACCEPTED MANUSCRIPT 13 It is also important to highlight the correlational nature of our results. Thus, while we have suggested that higher FC may index better information exchange between network regions engaged in interoceptive- and conceptualization-related processing, and that this could in turn promote higher EA, this remains just one of
RI PT
many possible causal interpretations of our correlational results. Future studies will be necessary to disentangle the directions of causal influence that may be involved (e.g., by examining changes in FC as a result of EA training). Finally, it should be emphasized that our hypotheses in this study were based on 1) the theoretical/empirical links between LEAS scores, conceptualization of emotions, and interoception
SC
(e.g., reviewed in Lane et al., 2015), and 2) the theoretical/empirical links between the latter two
psychological processes and the DMN and SN, respectively (e.g., reviewed in Barrett and Satpute, 2013;
M AN U
Lindquist and Barrett, 2012). We have therefore interpreted our neuroimaging results to support the link between LEAS scores and those psychological processes. However, because there does not appear to be a 1-to-1 mapping between brain regions and psychological processes (e.g., see Anderson, 2014), other interpretations of our results may also be possible. Future studies will therefore also be necessary to confirm that the associations we have observed between EA, the DMN, and the SN are best explained by
TE D
the conceptualization- and interoception-related processes that we have suggested.
In conclusion, this study found that greater EA, as measured by the LEAS, was associated with greater FC within the DMN and also within the SN. We have suggested this may index improved perceptual processing of bodily feelings, and improved conceptualization of those feelings in the context of an
EP
emotional reaction. Given that LEAS scores have been associated with several clinical phenomena associated with physical and emotional health, future research should examine these LEAS-FC
AC C
relationships in clinical populations. This could perhaps lead to novel, potentially predictive neural correlates of low EA and associated physical/mental health risk.
References
Anderson, M., 2014. After Phrenology: Neural Reuse and the Interactive Brain. MIT Press, Cambridge, MA.
Barchard, K., Bajgar, J., Leaf, D., Lane, R., 2010. Computer scoring of the Levels of Emotional Awareness
ACCEPTED MANUSCRIPT 14 Scale. Behav. Res. Methods 42, 586–595.
Barchard, K., Hakstian, A., 2004. The nature and measurement of emotional intelligence abilities; basic
Psychol. Meas. 64, 437–462.
RI PT
dimensions and their relationships with other cognitive abilities and personality variables. Educ.
Barrett, L., Lane, R., Sechrest, L., Schwartz, G., 2000. Sex Differences in Emotional Awareness. Personal. Soc. Psychol. Bull. 26, 1027–1035. doi:10.1177/01461672002611001
SC
Barrett, L., Satpute, A., 2013. Large-scale brain networks in affective and social neuroscience: towards an
doi:10.1016/j.conb.2012.12.012
M AN U
integrative functional architecture of the brain. Curr. Opin. Neurobiol. 23, 361–72.
Barrett, L., Simmons, W., 2015. Interoceptive predictions in the brain. Nat. Rev. Neurosci. 16, 419–29. doi:10.1038/nrn3950
Baslet, G., Termini, L., Herbener, E., 2009. Deficits in emotional awareness in schizophrenia and their
TE D
relationship with other measures of functioning. J. Nerv. Ment. Dis. 197, 655.
Behzadi, Y., Restom, K., Liau, J., Liu, T., 2007. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 37, 90–101.
EP
doi:10.1016/j.neuroimage.2007.04.042
Berthoz, S., Ouhayoun, B., Parage, N., 2000. Etude preliminaire des niveaux de conscience emotionnelle
AC C
chez des patients deprimes et des controles. (Preliminary study of the levels of emotional awareness in depressed patients and controls.). Ann. Med. Psychol. 158, 665–672.
Bréjard, V., Bonnet, A., Pedinielli, J., 2012. The role of temperament and emotional awareness in risk taking in adolescents. L’Encéphale Rev. Psychiatr. Clin. Biol. Thérapeutique 38, 1–9.
Buckner, R., 2013. The Cerebellum and Cognitive Function: 25 Years of Insight from Anatomy and Neuroimaging. Neuron 80, 807–815. doi:10.1016/j.neuron.2013.10.044
ACCEPTED MANUSCRIPT 15 Buckner, R., Andrews-Hanna, J.R., Schacter, D.L., 2008. The Brain’s Default Network. Ann. N. Y. Acad. Sci. 1124, 1–38. doi:10.1196/annals.1440.011
Bydlowski, S., Corcos, M., Jeammet, P., Paterniti, S., Berthoz, S., Laurier, C., Chambry, J., Consoli, S.,
RI PT
2005. Emotion-processing deficits in eating disorders. Int. J. Eat. Disord. 37, 321–329.
Chai, X., Nieto-Castanon, A., Öngür, D., Whitfield-Gabrieli, S., 2012. Anticorrelations in resting state networks without global signal regression. Neuroimage 59, 1420–1428.
SC
doi:10.1016/j.neuroimage.2011.08.048
Ciarrochi, J., Caputi, P., Mayer, J., 2003. The distinctiveness and utility of a measure of trait emotional
M AN U
awareness. Pers. Individ. Dif. 34, 1477–1490.
Consoli, S., Lemogne, C., Roch, B., Laurent, S., Plouin, P., Lane, R., 2010. Differences in emotion processing in patients with essential and secondary hypertension. Am. J. Hypertens. 23, 515–521.
Craig, A.D., 2009. How do you feel--now? The anterior insula and human awareness. Nat. Rev. Neurosci.
TE D
10, 59–70.
Critchley, H., 2005. Neural mechanisms of autonomic, affective, and cognitive integration. J. Comp.
EP
Neurol. 493, 154–66. doi:10.1002/cne.20749
Critchley, H., Mathias, C., Josephs, O., O’Doherty, J., Zanini, S., Dewar, B.-K., Cipolotti, L., Shallice, T., Dolan, R., 2003. Human cingulate cortex and autonomic control: converging neuroimaging and
AC C
clinical evidence. Brain 126, 2139–52. doi:10.1093/brain/awg216
Donges, U., Kersting, A., Dannlowski, U., Lalee-Mentzel, J., Arolt, V., Suslow, T., 2005. Reduced awareness of others’ emotions in unipolar depressed patients. J. Nerv. Ment. Dis. 193, 331–337.
Farb, N., Segal, Z., Anderson, A., 2013. Attentional modulation of primary interoceptive and exteroceptive cortices. Cereb. cortex 23, 114–26. doi:10.1093/cercor/bhr385
Frewen, P., Lane, R., Neufeld, R., Densmore, M., Stevens, T., Lanius, R., 2008. Neural correlates of levels
ACCEPTED MANUSCRIPT 16 of emotional awareness during trauma script-imagery in posttraumatic stress disorder. Psychosom. Med. 70, 27–31.
Gusnard, D., Akbudak, E., Shulman, G., Raichle, M., 2001. Medial prefrontal cortex and self-referential
RI PT
mental activity: relation to a default mode of brain function. Proc. Natl. Acad. Sci. 98, 4259–4264.
Kalisch, R., Wiech, K., Critchley, H.D., Dolan, R.J., 2006. Levels of appraisal: a medial prefrontal role in high-level appraisal of emotional material. Neuroimage 30, 1458–66.
SC
doi:10.1016/j.neuroimage.2005.11.011
La Rovere, M., Bigger, J., Marcus, F., Mortara, A., Schwartz, P., 1998. Baroreflex sensitivity and heart-rate
doi:10.1016/S0140-6736(97)11144-8
M AN U
variability in prediction of total cardiac mortality after myocardial infarction. Lancet 351, 478–484.
Lackner, J., 2005. Is IBS a problem of emotion dysregulation? Testing the levels of emotional awareness model, in: Presented at the Annual Meeting of the American Psychosomatic Society.
TE D
Lane, R., Carmichael, C., Reis, H., 2011. Differentiation in the momentary rating of somatic symptoms covaries with trait emotional awareness in patients at risk for sudden cardiac death. Psychosom. Med. 73, 185–92. doi:10.1097/PSY.0b013e318203b86a
EP
Lane, R., Quinlan, D., Schwartz, G., Walker, P., Zeitlin, S., 1990. The Levels of Emotional Awareness Scale: a cognitive-developmental measure of emotion. J. Pers. Assess. 55, 124–34.
AC C
doi:10.1080/00223891.1990.9674052
Lane, R., Reiman, E., Axelrod, B., Yun, L., Holmes, A., Schwartz, G., 1998a. Neural correlates of levels of emotional awareness. Evidence of an interaction between emotion and attention in the anterior cingulate cortex. J. Cogn. Neurosci. 10, 525–535.
Lane, R., Schwartz, G., 1987. Levels of emotional awareness: a cognitive-developmental theory and its application to psychopathology. Am. J. Psychiatry 144, 133–143.
ACCEPTED MANUSCRIPT 17 Lane, R., Sechrest, L., Reidel, R., Weldon, V., Kaszniak, A., Schwartz, G., 1996. Impaired verbal and nonverbal emotion recognition in alexithymia. Psychosom. Med. 58, 203–10.
Lane, R., Sechrest, L., Riedel, R., 1998b. Sociodemographic correlates of alexithymia. Compr. Psychiatry
RI PT
39, 377–385. doi:10.1016/S0010-440X(98)90051-7
Lane, R., Sechrest, L., Riedel, R., Shapiro, D., Kaszniak, A., 2000. Pervasive emotion recognition deficit common to alexithymia and the repressive coping style. Psychosom. Med. 62, 492–501.
SC
Lane, R., Weihs, K., Herring, A., Hishaw, A., Smith, R., 2015. Affective agnosia: Expansion of the alexithymia construct and a new opportunity to integrate and extend Freud’s legacy. Neurosci.
M AN U
Biobehav. Rev. 55, In Press. doi:10.1016/j.neubiorev.2015.06.007
Levine, D., Marziali, E., Hood, J., 1997. Emotion processing in borderline personality disorders. J. Nerv. Ment. Dis. 185, 240–246.
Li, W., Mai, X., Liu, C., 2014. The default mode network and social understanding of others: what do brain
TE D
connectivity studies tell us. Front. Hum. Neurosci. 8, 74. doi:10.3389/fnhum.2014.00074
Lindquist, K., Barrett, L., 2012. A functional architecture of the human brain: emerging insights from the
EP
science of emotion. Trends Cogn. Sci. 16, 533–540. doi:10.1016/j.tics.2012.09.005
McRae, K., Reiman, E., Fort, C., Chen, K., Lane, R., 2008. Association between trait emotional awareness and dorsal anterior cingulate activity during emotion is arousal-dependent. Neuroimage 41, 648–55.
AC C
doi:10.1016/j.neuroimage.2008.02.030
Medford, N., Critchley, H., 2010. Conjoint activity of anterior insular and anterior cingulate cortex: awareness and response. Brain Struct. Funct. 214, 535–549. doi:10.1007/s00429-010-0265-x
Moeller, S., Konova, A., Parvaz, M., Tomasi, D., Lane, R., Fort, C., Goldstein, R., RZ, G., 2014. Functional, Structural, and Emotional Correlates of Impaired Insight in Cocaine Addiction. JAMA Psychiatry 71, 61. doi:10.1001/jamapsychiatry.2013.2833
ACCEPTED MANUSCRIPT 18 Peelen, M., Atkinson, A., Vuilleumier, P., 2010. Supramodal representations of perceived emotions in the human brain. J. Neurosci. 51, 10127–34. doi:10.1523/JNEUROSCI.2161-10.2010
Raichle, M., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001. A default
RI PT
mode of brain function. Proc Natl Acad Sci U S A 98, 676–682. doi:10.1073/pnas.98.2.676
Roberton, T., Daffern, M., Bucks, R., 2013. Oral administration of the Levels of Emotional Awareness Scale. Aust. J. Psychol. 65, 172–179. doi:10.1111/ajpy.12018
SC
Roy, M., Shohamy, D., Wager, T.D., 2012. Ventromedial prefrontal-subcortical systems and the generation
10.1016/j.tics.2012.01.005
M AN U
of affective meaning. Trends Cogn. Sci. 16, 147–156. doi:S1364-6613(12)00027-7 [pii]
Shirer, W., Ryali, S., Rykhlevskaia, E., Menon, V., Greicius, M., 2012. Decoding subject-driven cognitive states with whole-brain connectivity patterns. Cereb. cortex 22, 158–65. doi:10.1093/cercor/bhr099
Smith, R., Braden, B., Chen, K., Ponce, F., Lane, R., Baxter, L., 2014a. The neural basis of attaining
TE D
conscious awareness of sad mood. Brain Imaging Behav. 9. doi:10.1007/s11682-014-9318-8
Smith, R., Fass, H., Lane, R., 2014b. Role of medial prefrontal cortex in representing one’s own subjective emotional responses: A preliminary study. Conscious. Cogn. 29, 117–130.
EP
doi:10.1016/j.concog.2014.08.002
Smith, R., Lane, R., 2015. The neural basis of one’s own conscious and unconscious emotional states.
AC C
Neurosci. Biobehav. Rev. 57, 1–29. doi:10.1016/j.neubiorev.2015.08.003
Subic-Wrana, A., Beetz, M., Paulussen, J., Wiltnik, J., Beutel, M., 2007. Relations between attachment, childhood trauma, and emotional awareness in psychosomatic inpatients, in: Presented at the Annual Meeting of the American Psychosomatic Society.
Subic-Wrana, C., Bruder, S., Thomas, W., Lane, R., Köhle, K., 2005. Emotional awareness deficits in inpatients of a psychosomatic ward: a comparison of two different measures of alexithymia.
ACCEPTED MANUSCRIPT 19 Psychosom. Med. 67, 483–489.
Tavares, P., Barnard, P., Lawrence, A., 2011. Emotional complexity and the neural representation of
RI PT
emotion in motion. Soc. Cogn. Affect. Neurosci. 6, 98–108. doi:10.1093/scan/nsq021
Thayer, J., Ahs, F., Fredrikson, M., Sollers, J.J., Wager, T.D., 2012. A meta-analysis of heart rate
variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health. Neurosci. Biobehav. Rev. 36, 747–56.
SC
Thayer, J., Lane, R., 2009. Claude Bernard and the heart-brain connection: further elaboration of a model
M AN U
of neurovisceral integration. Neurosci. Biobehav. Rev. 33, 81–88.
Verkuil, B., Brosschot, J., Tollenaar, M., Lane, R., Thayer, J., 2016. Prolonged Non-metabolic Heart Rate Variability Reduction as a Physiological Marker of Psychological Stress in Daily Life. Ann. Behav. Med. 1–11. doi:10.1007/s12160-016-9795-7
TE D
Wechsler, D., 2011. WASI -II: Wechsler abbreviated scale of intelligence - second edition, WASI.
Whitfield-Gabrieli, S., Nieto-Castanon, A., 2012. Conn: A functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect. 0, 125–41. doi:10.1089/brain.2012.0073
EP
Whitfield-Gabrieli, S., Thermenos, H., Milanovic, S., Tsuang, M., Faraone, S., McCarley, R., Shenton, M., Green, A., Nieto-Castanon, A., LaViolette, P., Wojcik, J., Gabrieli, J., Seidman, L., 2009. Hyperactivity and hyperconnectivity of the default network in schizophrenia and in first-degree
AC C
relatives of persons with schizophrenia 106, 1279–1284. doi:10.1073/pnas.0809141106
Zaki, J., Davis, J.I., Ochsner, K., 2012. Overlapping activity in anterior insula during interoception and emotional experience. Neuroimage 62, 493–499. doi:10.1016/j.neuroimage.2012.05.012
ACCEPTED MANUSCRIPT 20 Figure Legends
Figure 1. Visual depiction of some of the regions of interest (ROIs) within the DMN (Top; Red) and SN (Bottom; Anterior Network = Yellow, Posterior Network = Blue) that were used in the analyses described
RI PT
in the text. DMN ROIs: MPFC = medial prefrontal cortex (also covers portions of the anterior cingulate
cortex and orbitofrontal cortex), PCC = posterior cingulate cortex (also covers portions of the precuneus), MCC = mid-cingulate cortex, thal = thalamus, L Hc = left hippocampus, R Hc = right hippocampus. SN
ROIs: dACC = dorsal anterior cingulate cortex (also covers portions of dorsal MPFC and supplementary
SC
motor area), R AI = right anterior insula, L AI = left anterior insula, R PI = right posterior insula, L PI =
M AN U
left posterior insula.
Figure 2. Visual depiction of the regions within the DMN (left) and SN (right) that showed a significant positive correlation between functional connectivity and LEAS scores (after statistically accounting for age, gender, and IQ score). Seed regions are labeled in black; target regions are labeled in red. DMN ROIs: MPFC = medial prefrontal cortex (also covers portions of the anterior cingulate cortex and orbitofrontal cortex), PCC = posterior cingulate cortex (also covers portions of the precuneus), MCC = mid-cingulate
TE D
cortex, thal = thalamus, L AG = left angular gyrus, L Hc = left hippocampus, R Hc = right hippocampus. SN ROIs: dACC = dorsal anterior cingulate cortex (also covers portions of dorsal MPFC and supplementary motor area), R AI = right anterior insula, L PI = left posterior insula, L PreC = left
EP
precuneus, L SMG = left supramarginal gyrus (also covers portions of the inferior parietal gyrus), L cer (A) = left cerebellum (anterior network ROI), L cer (P) = left cerebellum (posterior network ROI), R cer (A) =
AC C
right cerebellum (anterior network ROI).
Figure 3. Example scatter plots illustrating the positive relationship between LEAS total scores and functional connectivity values within four ROI seed-target pairs (i.e., two from each network analysis). See Tables 1 and 2 for associated statistics (i.e., which also account for included covariates). MPFC = medial prefrontal cortex; PCC = posterior cingulate cortex; dACC = dorsal anterior cingulate cortex; R AI = right anterior insula.
ACCEPTED MANUSCRIPT
Table 1. Functional Connectivity Results within the Default Mode Network: Positive Correlations with LEAS Scores 1
RI PT
Table 1.1: Seed ROI = MPFC
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
PCC
2.90
p = .017
.82 (+/- .18)
Right Hippocampus
2.62
p = .025
M AN U
SC
Target ROI
Tscore
.26 (+/- .14)
One-Sample Ttest*
t = 23.14, p < .001
t = 9.62, p < .001
TE D
Left Angular Gyrus 2.38 p = .035 .57 (+/- .25) t = 11.62, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0). Table 1.2: Seed ROI = PCC
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
EP
Target ROI
Tscore
3.68
p = .011
.28 (+/- .14)
t = 10.51, p < .001
Bilateral Thalamus
2.95
p = .017
.04 (+/- .18)
t = 1.17, p = .26
AC C
Left Hippocampus
1Peak
Threshold = P < 0.001, Uncorrected; Cluster Threshold = P< 0.05, FDRCorrected
ACCEPTED MANUSCRIPT
2.90
p = .017
.82 (+/- .18)
Mid-Cingulate
2.20
p = .043
.46 (+/- .17)
t = 23.14, p < .001
RI PT
MPFC
t = 13.42, p < .001
SC
Right Hippocampus 2.15 p = .043 .23 (+/- .13) t = 8.97, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
Table 2.1: Seed ROI = dACC
M AN U
Table 2. Functional Connectivity Results within the Salience Network: Positive Correlations with LEAS Scores 2
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
Left Cerebellum (Anterior Salience Network)
3.82
p = .011
.32 (+/- .14)
t = 11.30, p < .001
TE D
Target ROI
Tscore
AC C
EP
Left Cerebellum (Posterior Salience Network) 3.80 p = .011 .10 (+/- .16) t = 3.06, p = .005 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0). Table 2.2: Seed ROI = right AI
Target ROI
2Peak
Tscore
p-value (FDRCorrected)
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
Threshold = P < 0.001, Uncorrected; Cluster Threshold = P< 0.05, FDRCorrected
ACCEPTED MANUSCRIPT
3.12
p = .039
.19 (+/- .23)
t = 4.16, p < .001
RI PT
Left Precuneus
Left SMG 3.11 p = .039 .44 (+/- .25) t = 9.17, p < .001 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
Tscore
Left Cerebellum (Anterior Salience Network)
4.91
p = .003
Mean (+/- SD) Connectivity Value (Z scores)
One-Sample Ttest*
.08 (+/- .13)
t = 3.04, p = .006
M AN U
Target ROI
p-value (FDRCorrected)
SC
Table 2.3: Seed ROI = left PI
AC C
EP
TE D
Right Cerebellum (Anterior Salience Network) 3.80 p = .011 .00 (+/- .12) t = .004, p = .996 *Comparing mean connectivity values to the null hypothesis (i.e., mean = 0).
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT