Association of Selective Attention and Functional Connectivity in Cerebral Visual Impairment

1814 

Abstract Submission 

Negin Nadvar¹, Claire Manley², Marie Drottar¹, Lotfi Merabet², Corinna Bauer^1,2

¹Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, ²Department of Ophthalmology, Massachusetts Eye and Ear, Harvard Medical School, Boston, MA

Negin Nadvar  
Department of Radiology, Massachusetts General Hospital, Harvard Medical School
Boston, MA

Claire Manley  
Department of Ophthalmology, Massachusetts Eye and Ear, Harvard Medical School
Boston, MA

Marie Drottar  
Department of Radiology, Massachusetts General Hospital, Harvard Medical School
Boston, MA

Lotfi Merabet  
Department of Ophthalmology, Massachusetts Eye and Ear, Harvard Medical School
Boston, MA

Corinna Bauer  
Department of Radiology, Massachusetts General Hospital, Harvard Medical School|Department of Ophthalmology, Massachusetts Eye and Ear, Harvard Medical School
Boston, MA|Boston, MA

Cerebral visual impairment (CVI) encompasses any visual deficit due to brain maldevelopment or damage and is a leading cause of visual disorders in the pediatric population in many parts of the world. Higher-order visual processes, such as visual search, are often impacted in individuals with CVI (Boot et al., 2010; Zhang et al., 2022) . A recent study investigated performance in a visual search task in individuals with CVI and found an overall impairment in time spent and accuracy in identifying the test target (Manley et al., 2023). However, the neural correlates of visual search tasks remain unknown in CVI. As a result, in the present study, we examined the association between performance in a selective attention visual task and resting-state functional connectivity (rsFC) in the brain networks involving attention and visual processing.

Sixteen sighted control (SC: 22.06 years ± 4.17 (sd), 10 females) and 10 CVI (17.70 years ± 5.29 (sd), 4 females) participants were included in the study. Participants completed a conjunction search task with mean performance accuracy and response time as the main outcomes of interest (Manley et al., 2023). For each participant, a T1W image (TR = 6.5 ms, TE = 2.9 ms, resolution = 1 mm³ isotropic), 2 resting-state fMRI (rsfMRI) runs (TR = 800 ms, TE = 30ms, duration = 5 min 15 sec, resolution = 2.25 x 2.25 x 2.4 mm³), and reverse phase-encoded field maps were acquired. Preprocessing steps included brain extraction, Mcflirt, b0 unwarping, 5mm smoothing, removal of motion components using ICA-AROMA, and regression of the white matter and CSF signal using FSL tools. Mean time series were extracted from 54 cortical regions derived from the Glasser atlas (Glasser et al., 2016) and 6 thalamic subdivisions from FreeSurfer (Iglesias et al., 2018). Four networks were considered: the dorsal attention network (DAN, 22 ROIs), ventral attention network (VAN, 26 ROIs), early visual network (VIS, 6 ROIs), and thalamus (Thal, 6 ROIs). Average rsFC was calculated within DAN, VAN, and VIS networks, between each of these networks and the rest of the networks under study (DAN, VAN, VIS, and Thal), as well as between each of the attention networks and Thal. For each of the inter-network and intra-network connections, the Spearman partial correlation was calculated between the task behavioral measures and rsFC across the subjects in each group, while controlling for the potential effects of age and verbal IQ scores. The results were corrected for multiple comparisons using False Discovery Rate (FDR) correction across all the inter- and intra-network study cases.

Our analysis found significant (p < 0.05 corrected) correlations between performance accuracy and rsFC within the DAN and VAN networks (r = 0.78, p = 0.007 and r = 0.72, p = 0.014, respectively) and between each of the DAN, VAN, and VIS networks and the rest of the networks under study (r = 0.65, p = 0.02, r = 0.67, p = 0.02 and r = 0.63, p = 0.02, respectively), for the SC group. The CVI group did not demonstrate any significant correlations between behavior and rsFC measures. However, the uncorrected results showed a trend for weak rsFC-behavior correlation in the negative direction for the CVI group for the DAN and VAN within-network and DAN, VAN, and VIS between-network analyses, while yielding a positive correlation for the SC group (Figure 1).

Overall, our results indicated a positive correlation between the rsFC involving attention networks and selective attention task in SC (i.e., higher rsFC in these networks is associated with more accurate performance). Our results provided evidence for an aberrant negative association for the CVI group. This change in rsFC-behavior association may be due to the neurological impairment in CVI which has important implications in rehabilitation. This notable hypothesis warrants further investigation in a larger number of individuals affected by CVI.

fMRI Connectivity and Network Modeling ¹

Attention: Visual ²

Computational Neuroscience

FUNCTIONAL MRI

Vision

Other - CVI, selective attention, resting-state functional connectivity