In my talk I will present a study on neural processing of voices. The aim was to investigate the possibility of using ERPs as a measure of recognition of a familiar voice. The methodology however raises questions concerning pre-attentive processing of voices. I will present the study on voice familiarity and discuss the typical MMN (Mismatch Negativity) that was found in relation to voices, but not to familiarization. Acoustic analysis of voice characteristics in the current study as well as follow up studies with controlled exposure and voice parameters will also be addressed. I would like to discuss these issues with you, and also the implications of a possible MMN to familiar voices.