Neuro-fuzzy classification of transcranial Doppler signals with chaotic meaures and spectral parameters

Transcranial Doppler (TCD) is a non-invasive diagnosis method which is used in diagnosis of various brain diseases by measuring the blood flow velocities in brain arteries. In this study, chaos analysis of the TCD signals recorded from the middle arteries of the temporal region of brain of the 82 patients and 23 healthy people was investigated. Among 82 patients, 20 of them had cerebral aneurism, 10 had brain hemorrhage, 22 had cerebral oedema and the remaining 30 had brain tumor. Maximum Lyapunov exponent which is the strongest quantitative indicator of chaos was found to be positive for all TCD signals. The correlation dimension was found as greater than 2 and as fractional value for all TCD signals. These two features were used for training a NEFCLASS model. The NEFCLASS model had two input nodes for D2 and maximum Lyapunov exponent values and five output nodes representing the subject group to which the inputs belonged. In order to make k-fold cross-validation, the data set was randomly divided into 5 subsets of equal size. In an iterated manner, 4 of these subsets were used for training and the remaining 1 subset was used for testing. This operation was repeated for 3 times. The average accuracy for train and test set was found as %81 and %79, respectively. The performance of the NEFCLASS model was also assessed in the same manner with spectral parameters (i.e. resistivity index and pulsatility index) which were obtained from Doppler sonograms. The average accuracy was found as %67 and %63 for train and test set, respectively.