Authors :
Presenting Author: Xiyu Feng, MSc – UCL Great Ormond Street Institute of Child Health
Hua Xie, PhD – Washington, D.C., USA, Children’s National Hospital; Freya Prentice, MS – London, UK, UCL Great Ormond Street Institute of Child Health; Priyanka Illapani, MS – Washington, D.C., USA, Children’s National Hospital; Rory Piper, MD – London, UK, UCL Great Ormond Street Institute of Child Health; Seok Jun Hong, PhD – Seoul, South Korea, Sungkyunkwan University; William Gaillard, MD – Washington, D.C., USA, Children’s National Hospital; Torsten Baldeweg, PhD – London, UK, UCL Great Ormond Street Institute of Child Health; Leigh Sepeta, PhD – Washington, D.C., USA, Children’s National Hospital
Rationale:
Thalamic connectivity alterations exist in patients with temporal lobe epilepsy (TLE) and may identify therapeutic neuromodulation targets (Piper et al., 2022) or guide surgical candicacy (He et al., 2017). Prior studies of TLE have found that functional connectivity of thalamus and predefined subparcellations affect seizure control (Chari et al., 2022; Jo et al., 2019). Here we employed a data-driven approach called ‘connectopic mapping’ to study thalamic functional organization without specific parcellations. Our aims were to investigate the following: 1) connectopic maps showing spatial patterns of thalamic function, 2) projection maps displaying thalamus connections to the whole brain, and 3) how these maps relate to clinical variables in children with TLE.
Methods:
We studied 65 children aged 5-18 with TLE (left TLE
n=50, right
n=13, bilateral
n=2) undergoing pre-operative verbal generation task fMRI at Great Ormond Street Hospital, London, UK. Connectopic Mapping (Haak et al. 2017). We correlated the timeseries of the left and right thalamus with the rest of the brain. We then calculated a within-thalamus similarity matrix of functional connectivity and applied non-linear manifold learning to this matrix, yielding connectopic maps (i.e., gradients) for left and right thalamus, respectively. We also computed projection maps (P maps) of the gradients to visualize how the thalamocortical functional connections vary throughout the entire brain. We used a SurfStat linear model to test whether clinical variables (age, gender, and laterality of seizure focus) affect thalamic-cortical P maps.
Results:
The primary connectopic map of the thalamus followed an anterior-to-posterior axis (Figure 1). The P map revealed that the anterior thalamus exhibited functional coupling with the prefrontal and orbitofrontal cortex and the anterior hippocampus. The mid-zone of the thalamus was coupled with the primary motor cortex. The posterior thalamus was coupled with the somatosensory and visual cortex and posterior hippocampus. The secondary connectopic map followed a dorsal-to-ventral axis (Figure 2), and the P map mirrored the geometric pattern of this dorsoventral secondary gradient. Both left (Pearson correlation
r=0.97-0.98,
p< 0.05) and right (
r=0.56-0.73,
p< 0.05) TLE patients displayed consistent primary and secondary P maps when compared to the overall TLE group. Thalamic-cortical P maps showed no significant associations with age, gender, or laterality of seizure focus, both for the TLE group overall and the left TLE subgroup (FDR-corrected
p >0.05).
Conclusions:
The study detected anterior-to-posterior and dorsal-to-ventral gradients within the thalami in children with TLE. These gradient projections showed the functional coupling patterns of thalamus with the whole brain, which were consistent across patients with left and right TLE, and unaffected by age and gender. Further research is needed to explore the clinical potential of applying connectopic mapping to thalamus in the context of epilepsy.
Funding:
K23 NINDSNS093152 to LNS NINDS Division of Intramural Research
NIH (P50 HD105328), DC-IDDRC, Human and Animal Imaging Core