Brain Tissue Classification by Its Ultrasonic Backscatter

A theoretical model to predict values of the backscatter coefficient for soft tissue was developed in terms of two parameters of inhomogeneous tissue. These parameters were the variances and the correlation lengths of fluctuations in mass density and bulk modulus. Then experimental measurements of backscatter coefficients were made on the brains of live calves. The frequency dependences of the experi- mentally determined backscatter coefficients were used to classify all the data into four groups. These four groups were found to correlate with four different types of brain tissue. Values of both parameters of the inhomogeneities for two of these four tissue types were chosen to give a good fit between theoretically predicted and experimentally measured backscatter coefficients. The correlation lengths found in this way seem to correspond closely to typical length scales seen in the histological sections for these two tissue types. The parameter vari- ances found in this way are reasonable, but no independent verification is available for them.