Research paper Statistical detection and analysis of mismatch negativity derived by a multi-deviant design from normal hearing children Miriam Petermann a , Peter Kummer b , Martin Burger a , Jörg Lohscheller a , Ulrich Eysholdt a , Michael Döllinger a, * a University Hospital Erlangen, Department of Phoniatrics and Paediatric Audiology, Medical School, Bohlenplatz 21, 91054 Erlangen, Bavaria, Germany b University Hospital München, Department of Phoniatrics and Paediatric Audiology, Medical School, Pettenkoferstrasse 4a, 80336 München, Germany article info Article history: Received 20 December 2007 Received in revised form 6 November 2008 Accepted 6 November 2008 Available online 27 November 2008 Keywords: Auditory evoked potentials (AEP) Mismatch negativity (MMN) Children Statistical analysis Multi-deviant paradigm abstract To shorten the time of measurement of mismatch negativity (MMN) and to objectify the analysis of MMN characteristics, auditory evoked potentials (AEP) were recorded and MMNs were statistically examined. AEPs from 16 healthy children between the ages of 5 and 7 were elicited using a multi-deviant design including: frequency, duration, intensity, and gap. For detection of MMN an automatic method was applied based on statistical analysis (t-tests) of AEP sweeps. The incidences found were compared to the incidences of MMNs detected by visual inspection of difference waveforms. MMN features such as peak amplitude, peak latency, onset latency, duration of MMN, and area under the curve were evaluated for MMN intervals identified by statistical analysis. MMN incidences identified by statistical analysis (44) reflected a consistency of 84% to visually identified (52) incidences of MMN. Scalp distribution of MMN was mainly fronto-central. Peak latencies of the duration MMN differed significantly from those of the other deviant MMNs. The multi-deviant paradigm was successfully applied for measuring children’s AEPs. The statistical approach proved applicable for analysing the data objectively. Ó 2008 Elsevier B.V. All rights reserved. 1. Introduction Auditory evoked potentials (AEP) reflect changes in the cortical electrical activity resulting from response to an acoustic event (Ors et al., 2002). Hence, they are an objective way of investigating auditory processing (McArthur and Bishop, 2005; Jansson-Verkasalo et al., 2003). The mismatch negativity (MMN) is a late, endogenous compo- nent of the AEP and reflects specific auditory stimulus discrimina- tion processes of the human brain (Näätänen et al., 1978). According to Holopainen et al. (1997), the MMN is easy to record and convenient for children. The MMN is thought to be associated with automatic, subcon- scious memory processes. Repetition of a standard stimulus causes a trace in the auditory sensory memory (Näätänen et al., 2005). Any new incoming stimulus is compared to the created memory trace. If the incoming stimulus differs from the sequence of previ- ous stimuli, the MMN is generated. Hence, the MMN is the neuro- nal correlate of an automatic change-detection process (Näätänen et al., 2005), that is independent from attention. A sensory-specific process located in the auditory cortex is suggested to detect the deviant stimulus which then triggers a frontal process related to a passive attention switch (Giard et al., 1990). The deviant stimulus must differ from the standard stimulus in at least one stimulus attribute. In the simplest case, the standard stimulus is a pure tone and the deviant differs in frequency, dura- tion or intensity. In addition to sinusoidal tones (e.g., Korpilahti and Lang, 1994; Holopainen et al., 1997; Shafer et al., 2000; Morr et al., 2002; Lachmann et al., 2005; Kujala et al., 2006), speech stimuli such as vowels (e.g., Cheour-Luhtanen et al., 1995; Shestak- ova et al., 2003), syllables (e.g., Kraus et al., 1996; McGee et al., 1997; Uwer et al., 2002; C ˇ eponiené et al., 2004; Lachmann et al., 2005; Zeftawi, 2005) or pseudo-words (Sussman et al., 2004) are used. MMN is a negative half cycle waveform in the difference signal between the cortical responses to the deviant and the standard stimuli. It occurs around 200 ms after stimulus onset (McGee et al., 1997) in adults and somewhat later in children (Cheour et al., 1998; Shafer et al., 2000). The magnitude of the MMN corre- lates with the degree of difference in the standard and the deviant stimulus (Ponton et al., 1997; Picton et al., 2000). The better a sub- ject is able to distinguish the deviant from the standard stimuli, the larger the MMN (Novitski et al., 2004). Under the same experimental conditions in which MMN occurs, it is possible to observe a second negative component in the differ- ence signal (Alonso-Búa et al., 2006), known as late discriminative 0378-5955/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.heares.2008.11.001 * Corresponding author. Tel.: +49 9131 85 33814; fax: +49 9131 85 39272. E-mail addresses: Miriam.ct@gmx.de (M. Petermann), Peter.Kummer@med.uni- muenchen.de (P. Kummer), Burger.Martin@vdi.de (M. Burger), Joerg.Lohscheller@ uk-erlangen.de (J. Lohscheller), Ulrich.Eysholdt@uk-erlangen.de (U. Eysholdt), Michael.Doellinger@uk-erlangen.de (M. Döllinger). Hearing Research 247 (2009) 128–136 Contents lists available at ScienceDirect Hearing Research journal homepage: www.elsevier.com/locate/heares