Spoken word onset syllables (prime fragments) have been used to track neurophysiological processing of following written words (targets). Between 300 and 400 ms event-related potentials (ERPs) over the left hemisphere were more positive for targets that did not match their preceding prime fragments (e.g., hun-dragon) compared to matching targets (e.g., dra-dragon). This P350 effect has been related to the activation of modality independent neural word form representations. In the present experiment we set out to characterize neural word processing specific to the auditory domain. Spoken word onset syllables (prime fragments) were followed by spoken words (targets). Reduced amplitudes for matching targets were found for the N100 and the T-complex (100-300 ms), for the P350 (300-400 ms) and for a central negativity starting at 300 ms. The early potentials possibly index the priming of speech sound processing. The P350 replicates previous work with written words. This constitutes further evidence for shared neural word form representations in auditory and visual word recognition. The central negativity might be related to the rapid phonological matching of prime and target; or to the immediate testing of phonological expectations in speech recognition.