Developing screening method of autism using fractal analysis

本研究的目的是将人脑形态学的分形分析应用于自闭症大脑的有效筛选。从临床和社区样本中招募广泛性发育障碍（PDD）个体和典型发育对照。然后将原始MR图像转换为BMP轴向图像，并使用专门用于成像数据分形分析的Benoit for MatlabR进行分析。计算三维分形维数（FD）作为估计皮质褶皱复杂性和/或皮质体积的指标。正如预测的那样，两个诊断组在大脑和小脑FD上都是不同的。与对照组相比，PDD患者的脑FD更小，小脑FD更大，提示PDD患者神经发育异常或代偿，并可能在该领域应用分形分析。

The aim of the present study was to apply fractal analysis of the human brain morphology to efficient screening of the autistic brains. Individuals with pervasive developmental disorders(PDD) and controls with typical development were recruited from clinic and community samples. Original MR images were then converted to BMP axial images and analyzed with Benoit for MatlabR specialized for fractal analysis of the imaging data. Three-dimensional fractal dimension(FD) was calculated as an index for estimating the complexity of the cortical folding and/or cortical volume. As predicted, two diagnostic groups were different both in cerebral and cerebellar FD. Compared to the controls, the PDD had smaller cerebral FD and larger cerebellar FD, suggesting aberrant or compensatory neurodevelopment in PDD and possible clinical use of the fractal analysis in this field.