Dynamic functional connectivity analysis: uncovering disorder signatures in anorexia nervosa
Poster No:
475
Submission Type:
Abstract Submission
Authors:
Monica Di Giuliano1, Feliberto de la Cruz2, Andy Schumann1, Karl- Jürgen Bär1, Katrin Rieger1
Institutions:
1Klinikum Universität, Jena, Thuringia, 2Jena University Hospital, Jena, Germany
First Author:
Co-Author(s):
Introduction:
The maintenance and non-stationary switching between brain states is crucial for self-regulation, cognitive and emotional processes. These states represent stable neural conditions persisting over specific periods and have been studied to gain insights into neurological and neuropsychiatric disorders, utilizing methods like dynamic functional connectivity analysis (dFC). Nevertheless, the potential benefits of analyzing dFC have not been fully dived into regarding anorexia nervosa (AN), a severe and prominent mental disorder primarily affecting adolescent girls and young women, with the highest mortality rate among all psychiatric conditions. Underlying psychological and biological mechanisms driving dFC in AN - at whole brain level - are still to be elucidated. In this work our goal is to investigate the whole brain dFC in patients with AN and their relationship with disorder-defining characteristics, such as body mass index.
Methods:
We collected resting state fMRI data from 18 acute AN patients and 22 healthy controls (HCs), matched by age and gender. The sequence parameters were: TR = 484 ms, TE = 30 ms, voxels = 2.5 mm3, 1,900 whole brain volume sets. Data preprocessing was performed using the 'afni_proc.py' script. We extracted the average time series of 300 regions-of-interest exploiting a well-validated parcellation model [1], on the whole brain level. We adopted a sliding-window approach with 60 s window length and the onset of each window progressively shifted by 20 s (~40 TR) from that of the previous window (total number of windows = 44). In each window, we computed the Pearson's correlation between all parcellated brain regions, generating a symmetric FC matrix, whose values were then transformed to z score. All dFC matrices were concatenated across all participants, and k-means clustering was performed to identify a set of brain states. The elbow method was used to identify the optimal number of kernels: this procedure results in five cluster centroids used to classify the dFC matrices for each subject. For each brain state, we computed different dynamic features (dwell time, flexibility index, fraction of time spent).
Results:
Among the five identified brain functional states in both HCs and AN, state 3 stands out as clinically relevant in AN (Fig.1A, left). The time spent for patients and HCs was significantly different (p < 0.01) in state 3. We also observed a trend (p < 0.1) for the dwell time in state 3, with AN patients spending approximately 5 consecutive windows, compared to 4 for HCs. This state exhibited a strong coupling between ventral attention (VAN) – somotomotor (SMN) networks as well as by a weak coupling between dorsal attention with default mode and limbic networks (Fig.1A, right). Another finding was the robust correlation (p = 0.003) between body mass index and the time spent in state 3 (Fig. 2B), suggesting potential clinical significance for this specific brain state.
Conclusions:
The VAN system is salient for alteration of bodily signalling processing as well as low interoception, highlighting the possibility of a role in disconnection of large brain networks during self-representation and environmental salience processing [2]. Accordingly, the SMN is important for body image disturbances, as well as responsible for the failure of integrating visual and somatosensory perceptual information. Given the functional connectivity pattern between the VAN and SMN networks during state 3, we suggest that patients are less able to redirect and shift their focus from body-related stimuli to non-bodily-relate environmental stimuli. Therefore, brain states allow us to draw a peculiar therapeutical and interventional attention on the role of attentional and interoceptive bodily mechanisms in shaping and maintaining AN psychopathology. Notably, these mechanisms can be uncovered during dynamic functional connectivity changes in the whole brain, which appear to be less plastic and disrupted in AN.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Emotion, Motivation and Social Neuroscience:
Emotional Perception
Modeling and Analysis Methods:
fMRI Connectivity and Network Modeling
Task-Independent and Resting-State Analysis 2
Perception, Attention and Motor Behavior:
Perception and Attention Other
Keywords:
Eating Disorders
FUNCTIONAL MRI
Psychiatric Disorders
Other - Anorexia Nervosa
1|2Indicates the priority used for review
·dFC results
Provide references using author date format
[2] Kim, D. J., Moussa‐Tooks, A. B., Bolbecker, A. R., Apthorp, D., Newman, S. D., O'Donnell, B. F., & Hetrick, W. P. (2020). Cerebellar–cortical dysconnectivity in resting‐state associated with sensorimotor tasks in schizophrenia. Human Brain Mapping, 41(11), 3119-3132.