Purkinje (P) neurons are highly irregular in their discharge patterns having a high coefficient of variation. We have developed a method to extract the signal from these noisy neurons employing multiple linear regression against the input signals that are causal in P-cell firing, namely retinal slip, head motion parameters, and the efference copy of the oculomotor commands. This method was shown to extract all of the experimental stimulus-related signal as the residual following extraction had the same amplitude distribution and spectral content as those of the spontaneous discharge of P-cells when no external stimuli were applied. The method enables us to decompose P-cell firing modulation into components mediated by mossy fiber inputs to flocculus, that is, retinal slip, vestibular and oculomotor efference copy signals. By evaluating these components in relation to VOR adaptation, one can argue that one role of the flocculus is in converting these sensory and efference copy input signals into eye movement-related signals in conjunction with VOR motor learning.