24 subjects were included in the final analysis [Simple FS - 11 and Complex FS -13] within 12 days [IQR= 8.5 -13.5] and 10 days [IQR = 9-30] after the last seizure respectively. Among the clinical variables analysed, namely mean age at onset, a number of recurrent febrile seizures, the maximum duration of febrile seizures, duration of the disease, none showed significant between-group differences. Functional connectivity (FC) was estimated using time series correlation of seed region–to-whole-brain-voxels and the brain regions showing significant between group connectivity differences were correlated with disease-defining clinical characteristics. Our results revealed that CFS group had altered temporal lobe connectivity proportional to recurrences, and duration of the seizure.
Altered brain connectivity in complex febrile seizure
Seed to Voxel analysis revealed that the patterns of connectivity of multiple seed ROI’s in patients in the complex febrile seizure group were significantly different (FDR-corrected, P < 0.005) from those in simple febrile seizure group (Figure.1)(Table 2). Patients with CFS demonstrated increased connectivity involving bilateral temporal lobes involving the medial and lateral temporal cortices, bilateral thalamus, bilateral accumbens, bilateral amygdala and left supplementary cortex when compared to simple febrile seizure group. On the other hand, decreased functional connectivity of the primary sensory areas and Default Mode Network (DMN) was seen involving the Posterior Cingulate Cortex (PCC), bilateral postcentral gyrus, and left occipital lobe in the CFS group.
Increased connectivity in complex febrile seizure
Bilateral middle temporal gyrus revealed increased functional connectivity with the bilateral hippocampus, bilateral parahippocampal gyrus and bilateral accumbens. Right middle temporal gyrus (Figure 1a) had additional increased connectivity with bilateral thalamus and ipsilateral amygdala and putamen. Left middle temporal gyrus (Figure 1b) demonstrated increased connectivity involving the amygdala bilaterally, ipsilateral orbitofrontal and subcallosal cortex. The left supplementary cortex revealed increased connectivity with the parahippocampal gyrus and bilateral thalamus.
Decreased connectivity in complex febrile seizure
Precuneus and posterior cingulate cortex (Figure 1f) revealed decreased connectivity with bilateral supplementary motor cortex, postcentral gyrus, bilateral caudate and right lentiform nucleus. Left supplementary motor cortex (Figure 1e), in addition, showed decreased connectivity with left inferior frontal gyrus. Bilateral postcentral gyri (Figure 1 c,d) showed additional decreased connectivity with left lingual gyrus, fusiform gyrus and intracalcarine cortex.
Altered connectivity proportional to the seizure recurrence and duration
The number of recurrences of febrile seizures was positively correlated with brain functional connectivity of bilateral middle temporal gyri to right Hippocampus [right: r = 0.6; p = 0.003; left: r = 0.54, p = 0.0086], and Left Middle Temporal gyrus with right Parahippocampus [r = 0.57, p = 0.0055] and negatively correlated with left Supplementary Motor to Precuneus [r = -.043, p = 0.041] (Figure 2; Table 3).
Furthermore, duration of the longest febrile seizures was positively correlated (Figure 3) with the functional connectivity of right Middle Temporal gyrus (MTG) to left Hippocampus [r= 0.59, p = 0.003], right Parahippocampus [r = 0.58, p = 0.004], bilateral thalami [right, r= 0.65, p =0.001; left, r= 0.59, p= 0.003], right putamen [r = 0.60, p =0.002], amygdala [r = 0.59, p =0.003], bilateral accumbens [r= 0.49, p = 0.019] and left supplementary motor cortex to right Parahippocampus [r= 0.44, p = 0.04]. Additionally, duration of longest febrile seizures was negatively correlated with connectivity of the bilateral post central gyrus to left lingual gyrus (right: r= -0.48, p = 0.023; left: r= -.55, p =0.006), Precuneus (right : r= -.45, p =0.03; left r= -0.49, p = 0.02; ) and right post central gyrus to left intracalcarine cortex [ r= -0.53, p = 0.01] was noted (Fi 3; Table 3).
No statistically significant correlations of age of onset of illness, time interval between imaging and last seizures with connectivity changes were noted.