A SHAPE-BASED FINITE IMPULSE RESPONSE MODEL FOR FUNCTIONAL BRAIN IMAGES

We present a new finite impulse response (FIR) model for hemodynamics in functional brain images. Like other FIR models, our method permits a flexible formulation of the hemodynamic response. The distinctive feature of this model is that the shape information of the canonical HRF is imposed on the FIR, with little loss of flexibility. Model fitting involves minimizing quadratic error under linear constraints. Our model proves more robust to noise on both synthetic data and real data, in contrast to other FIR models.