Deconvolution of 40 Hz steady-state fields reveals two overlapping source activities of the human auditory cortex

Abstract

Steady-state auditory evoked fields were recorded from 15 subjects using a whole head MEG system. Stimuli were 800 ms trains of binaural clicks with constant stimulus onset asynchrony (SOA). Seven different SOA settings (19, 21, 23, 25, 27, 29 and 31 ms) were used to give click rates near 40 Hz.

Transient responses to each click were reconstructed using a new algorithm that deconvoluted the averaged responses to the different trains. Spatio-temporal multiple dipole modelling in relation to 3D MRI scans revealed two overlapping source components in both the left and right auditory cortex. The primary sources in the medial part of Heschl's gyrus exhibited a N19-P30-N40 m pattern. The secondary, weaker sources at more lateral sites on Heschl's gyrus showed a N24-P36-N46 m pattern. When applied to transient middle latency auditory evoked fields (MAEFs) recorded at SOAs of 95-135 ms, the primary sources imaged activities similar to the deconvoluted steady-state responses, but the secondary source activities were inconsistent.

Linear summation of the deconvoluted source waveforms accounted for more than 96% of the steady-state variance. This indicates that the primary activity of the auditory cortex remains constant up to high stimulation rates and is not specifically enhanced around 40 Hz.

Introduction

To process sounds and speech in everyday live, the auditory system must be able to estimate temporal patterns precisely and rapidly. Processing at the level of the auditory cortex can be studied non-invasively with auditory evoked potentials (AEP) or magnetic fields (AEF). However, due to overlap, response waveforms may cancel or enhance each other and form steady-state responses when the interval between the stimuli is shorter than the duration of the response (Regan, 1965, Regan, 1982). Galambos et al. (1981) proposed that a periodic 40 Hz auditory event-related potential resulted from the superimposition of successive deflections in the middle latency evoked potential (MAEP). This hypothesis has been supported by synthesis of 40 Hz steady-state response from transient MAEPs and middle latency auditory evoked fields (MAEF) recorded with stimulus rates around 10 Hz (Stapells et al., 1988, Hari et al., 1989, Plourde et al., 1991).

In the first report on auditory MAEPs, Geisler et al. (1958) suggested that they originated in the auditory cortex. Later studies of the scalp topography of the MAEP and of the responses recorded in patients with temporal lobe lesions remained controversial (Vaughan and Ritter, 1970, Picton et al., 1974, Cohen, 1982, Kraus et al., 1982, Wood and Wolpaw, 1982, Woods et al., 1987) although complete abolition of the N_a–P_a component of the MAEP occurred in one patient with bilateral lesions of the auditory cortex (Özdamar et al., 1982).

The predominantly cortical origin of the MAEP has been unambiguously demonstrated by dipole source analysis in patients with lesions, intracranial measurements, and magnetoencephalography (MEG). Scherg and von Cramon (1986) presented a spatio-temporal dipole model to explain the coronal scalp distribution of the MAEP with two dipole sources in each temporal cortex. The response recorded at the vertex showed three main waves: N_a at about 18 ms, P_a at about 30 ms and N_b at about 40 ms, but the response recorded from more lateral electrodes had different peak latencies. The spatio-temporal model separated a primary N9–P30 source component of mainly tangential orientation from a secondary N27–P39 component of more radial orientation. In more than 60 patients with circumscribed vascular temporal lobe lesions affecting either the auditory cortex or the thalamo-cortical auditory radiation, the MAEP source components were reduced or abolished in the lesioned hemisphere, but unaffected in the intact hemisphere (Scherg and von Cramon, 1986, Scherg and von Cramon, 1990).

Consistent with these findings, human intracranial recordings with depth (Celesia, 1976, Liegeois-Chauvel et al., 1991, Liegeois-Chauvel et al., 1994) and chronic subdural electrodes (Lee et al., 1984) have demonstrated early cortical responses to auditory stimuli similar to the N19–P30 complex. These responses were recorded from a restricted region on the medial portion of Heschl's gyrus. This area of koniocortex represents the primary auditory cortex (Galaburda and Sanides, 1980).

Confirming results have also been obtained by MEG dipole localisation of middle latency transient (stimulation rates up to 10 Hz) and steady-state fields (periodic stimulation above 10 Hz). Transient MAEFs showed various peaks with consistent source localisation in the superior temporal plane: N19/P30/P50 m (Scherg et al., 1989a, Mäkelä et al., 1994, Yoshiura et al., 1996, Huotilainen et al., 1998) with P30 m as the most prominent peak (Pelizzone et al., 1987, Pantev et al., 1995). Auditory evoked steady-state fields evoked by amplitude modulated (32 Hz) sinusoidal tones with different carrier frequencies showed that the human auditory cortex was tonotopically organised with the lower frequencies more lateral (Romani et al., 1982). Steady-state MAEFs elicited by 40 Hz clicks in trains (Mäkelä and Hari, 1987, Forss et al., 1993) or continuously (Hari et al., 1989) had at least one main source in the superior temporal cortex.

Makeig (1990) reported time-locked MAEP activity in the 40 Hz range in response to clicks presented at slow rates (<1 Hz). This activity was superimposed on the late N 100 component and separated by using a narrow band-pass filter. The author speculated that clicks presented at these slow rates activated, in addition to the MAEP peaks, one or more oscillatory response generators. MAEFs elicited by a 500 ms toneburst of 1000 Hz and filtered from 28–48 Hz showed a similar oscillatory response named gamma-band response (Pantev et al., 1991, Galambos, 1991). Oscillatory coupling between thalamic and auditory cortex has been suggested to generate both the gamma-band response and the 40 Hz steady-state field (Ribary et al., 1991, Llinas and Ribary, 1991). Galambos (1991) proposed the gamma-band response as the underlying physiological activity for both the transient and the steady state MAEP. Based on single dipole MEG localisations, however, Pantev et al. (1993) suggested that the gamma-band response was generated in a different cortical region from the transient and steady-state MAEF. These transient and steady-state responses were themselves dissociated on the basis of single dipole tonotopical organisation, with the transient MAEFs to lower frequencies being more medial rather than lateral (Pantev et al., 1995, Pantev et al., 1996).

The present study separated the 40 Hz steady-state response into its underlying physiological components and identified their related generators in the brain. To achieve this goal, we developed a new algorithm to deconvolute the steady-state responses recorded at several stimulus rates into model transient responses. A paradigm by Mäkelä and Hari (1987) that employed 40 Hz click trains to present trains was modified to use different inter-click intervals (19–31 ms) in the 40 Hz range.

According to previous studies (Hari et al., 1989, Picton et al., 1992), our first hypothesis was that the activity of the primary auditory cortex, as reflected in the N19–P30 complex, should not become either refractory or resonant at stimulation rates between 30 and 60 Hz. In this view the enhanced amplitude at rates near 40 Hz is caused by the superimposition of overlapping transient responses and not by any increased responsiveness at resonant frequencies. We therefore expected that (a) the steady-state response between 30 and 60 Hz can be linearly decomposed into a transient response, and (b) the deconvoluted response is highly similar to the transient response obtained at slow click intervals between 95–135 ms (about 9 Hz).

Based on our previous studies on multiple sources of the MAEP (Scherg and von Cramon, 1986, Scherg and von Cramon, 1990) and MAEF (Scherg et al., 1989a), we further hypothesised that at least two generating areas in the auditory cortex contribute to the steady-state response. Therefore, spatio-temporal multiple dipole source analysis (Scherg, 1990) was applied to the deconvoluted steady-state responses and source locations were matched with MRI scans.