Network analysis reveals abnormal functional brain circuitry in anxious dogs

**Background:** Anxiety disorders are common in both humans and dogs, with human neuroimaging studies showing abnormal functional brain networks. Dogs may serve as a natural translational model for human anxiety due to similarities in brain circuitry. This study aimed to investigate whether anxious dogs exhibit similar aberrant functional brain network topology using resting-state fMRI and graph theory.

**Methods:** The study included 25 healthy beagle dogs (6 castrated males, 19 neutered females; mean age 47±23 months for females, 64±32 months for males) and 13 patient dogs of various breeds (mean age not clearly reported for patients; individual ages ranged from 39 to 151 months). Anxiety symptoms were assessed using the Canine Behavioral Assessment & Research Questionnaire (C-BARQ). All dogs underwent rs-fMRI on a 3T Siemens Trio Tim scanner under anesthesia (dexmedetomidine premedication, propofol induction, isoflurane maintenance). A 30-region parcellated atlas was used to construct functional brain networks. Graph theoretical analysis was performed using GRETNA, calculating global and nodal network metrics (degree, characteristic path length, global efficiency, clustering coefficient, local efficiency, small-worldness) across network densities from 20% to 50%. Group differences were assessed using Mann-Whitney U tests with FDR correction. Spearman's rank correlations were computed between network metrics and C-BARQ scores.

**Key Results:** No significant differences were found in global network topology between groups. However, within the anxiety circuit (amygdala, frontal lobe, hippocampus, mesencephalon, thalamus), the patient group showed significantly increased global efficiency (p=0.007, FDR corrected) and local efficiency (p=0.003, FDR corrected) in the amygdala, and significantly decreased characteristic path length (p=0.006, FDR corrected) in the amygdala. Functional connectivity was significantly higher in patients for connections: amygdala-hippocampus (U=77, p=0.022), amygdala-mesencephalon (U=87, p=0.040), amygdala-thalamus (U=81, p=0.026), frontal lobe-hippocampus (U=64, p=0.009), frontal lobe-thalamus (U=78, p=0.028), and hippocampus-thalamus (U=52.5, p=0.005). Connectivity between hippocampus and mesencephalon was significantly lower in patients (U=66, p=0.014). Correlations with C-BARQ scores revealed: amygdala degree (ρ=0.594, p=0.049) and global efficiency (ρ=0.583, p=0.047) positively correlated with stranger-directed fear; amygdala degree (ρ=-0.593, p=0.042) and amygdala-amygdala connectivity (ρ=-0.781, p=0.042) negatively correlated with excitability; hippocampus degree (ρ=0.616, p=0.033) and amygdala-hippocampus (ρ=0.758, p=0.028) and amygdala-thalamus (ρ=0.905, p<0.001) connectivity positively correlated with attachment/attention-seeking; hippocampus degree positively correlated with trainability (ρ=0.579, p=0.049) and touch sensitivity (ρ=0.837, p=0.001); frontal lobe clustering coefficient positively correlated with chasing (ρ=0.784, p=0.003); frontal lobe characteristic path length negatively correlated with familiar dog aggression (ρ=-0.879, p=0.009), while frontal lobe local efficiency (ρ=0.805, p=0.029), global efficiency (ρ=0.879, p=0.009), and degree (ρ=0.954, p=0.001) positively correlated with familiar dog aggression.

**Clinical Implications:** These findings demonstrate that anxious dogs exhibit altered functional brain network topology within the anxiety circuit, particularly involving the amygdala, hippocampus, and frontal lobe, with specific connectivity changes correlating with behavioral symptoms. The results suggest that rs-fMRI could serve as a diagnostic tool for canine anxiety and potentially monitor treatment response. The study supports the use of dogs as a translational model for human anxiety disorders, though larger sample sizes are needed to confirm these observations.