Lateralisation Matters: Discrimination of TLE and MCI Based on SPHARM Description of Hippocampal Shape

2018 IEEE 31st International Symposium on Computer-Based Medical Systems (CBMS)(2018)

引用 1|浏览2
暂无评分
摘要
Spherical Harmonic coefficients computed from human and automated hippocampus segmentations are used to discriminate patients suffering from MCI, TLE, and healthy individuals. In particular, we consider unilateral TLE and investigate which of the two hippocampi should preferably be used for reliable classifications. We find that the preferable classification setup strongly depends on the actual discrimination task. Qualification of human raters turns out to be crucial for discriminating patients with left lateralised TLE (TLE-L) from healthy controls and MCI patients (while automated segmentation techniques entirely fail in these two tasks). For discriminating TLE-R patients from healthy controls and TLE-L, automated segmentation tools, especially when fusing several ones, turn out to be a good choice. With respect to lateralisation, some results are well in accordance with expectations derived from studies using different approaches (e.g. better MCI and TLE-L discrimination obtained on the left hippocampus). Results deviating from those expectations may be explained by compensation effects as observed in case one hemisphere takes over responsibilities from the pathological one, thereby also affecting its own structure.
更多
查看译文
关键词
Discrimination of neurophysiological diseases, SPHARM, (automated) hippocampus segmentation, MCI, TLE, lateralisation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要