COGNITIVE NEUROSCIENCE AND NEUROPSYCHOLOGY NEUROREPORT 0959-4965 & Lippincott Williams & Wilkins Vol 12 No 15 29 October 2001 3189 An electrophysiological response to different pitch contours in words Claudia K. Friedrich, CA Kai Alter and Sonja A. Kotz Max Planck Institute of Cognitive Neuroscience, Stephanstrasse 1a, D-04103 Leipzig, Germany CA Corresponding Author Received 17 July 2001; accepted 7 August 2001 A spoken word with more than one syllable contains a speci®c stress pattern found to be processed during spoken word recognition. The present study investigated the word's pitch contour as a single auditory parameter that marks stress. Event-related brain potentials (ERPs) were recorded while subjects made decisions to arti®cially pitch manipulated words. ERPs revealed that pitch contours are discriminated already within the ®rst syllable of a word. Furthermore, behavioral responses for words with incorrect pitch contours were longer than for words with correct pitch contours. The results suggest that the pitch contour is an auditory feature of the spoken word that a listener automatically processes during spoken word recognition. NeuroReport 12:3189±3191 & 2001 Lippincott Williams & Wilkins. Key words: Arti®cial speech; Auditory event-related brain potentials; Prosody; Word recognition INTRODUCTION A spoken word not only yields information about its constituting phonological segments, but also about a spe- ci®c stress pattern, realized via suprasegmental parameters such as pitch contour, duration and amplitude. Recent behavioral studies revealed that the stress information is processed during spoken word recognition in natural speech [1±4]. The aim of the present study was to investigate the brain response to variations in a single suprasegmental parameter. Two recent ERP studies presented naturally spoken bi-syllabic words with stress on the ®rst syllable (initi- ally stressed words) and with stress on the second syllable (initially unstressed words) [5,6]. An enhanced left-frontal distributed negativity at around 325 ms (N325) was reported when initially unstressed words with a reduced vowel (schwa) in the ®rst syllable were compared to initially stressed words. The N325 was interpreted as a marker of stress [5]. In contrast, no ERP difference for bisyllabic words with different stress pattern was found if initially unstressed words contained full vowels [6]. This suggests that the N325 might be correlated to vowel quality rather than to stress. As one of the suprasegmental parameters is the duration of a syllable, the detection of a stress-related ERP component could be counteracted by latency jitter. Such effects can be prevented when temporal characteristics of the to-be- compared stimuli are equal. Therefore, in the present study a single suprasegmental parameter was arti®cially manipulated by leaving the time course of the word constant. The pitch contour was chosen as it is sug- gested to be an important parameter of stress [7]. The study aimed to explore whether different pitch contours can be discriminated by the brain and if pitch informa- tion is processed in spoken word recognition. MATERIALS AND METHODS Subjects: Twenty-four right-handed students with normal hearing (12 female, age 18±30 years) participated in the study. Stimuli and procedure: Eighty bisyllabic German words (40 initially stressed, 40 initially unstressed) were pre- sented. Forty of the words belonged to the category animate, 40 to the category inanimate. The words were spoken by a female native speaker of German. The ®rst syllable of the words had a mean duration of 272 ms, the second syllable was on average 405 ms long. A speech editor (PRAAT version 3.2, Paul Boersma and David Weenink) was used to extract the F0 contour of one initially stressed (AMboss [anvil]) and one initially unstressed word (abTEI [abbey]). On the basis of both pitch contours, two versions of each word were resynthesized with an auto- matized procedure provided by PRAAT. Periodical parts of the signal were adjusted to realize one version of each word with an initially stressed, and another version of the same word with an initially unstressed pitch contour (Fig. 1). All presented words were digitally manipulated in their pitch contour as there is evidence that pitch modulation per se evokes an early response of the brain [8]. Subjects evaluated the stress pattern (correct vs incor- rect) in one task and made animacy judgments in another task. Stimulus presentation in both tasks was identical. Words were presented auditory via loudspeakers. A ®xa- tion cross in the center of the screen appeared 500 ms before stimulus presentation and remained on the screen