- Email: [email protected]
Development of the 40 Hz steady state auditory evoked magnetic field from ages 5 to 52
Clinical Neurophysiology 117 (2006) 110–117 www.elsevier.com/locate/clinph
Development of the 40 Hz steady state auditory evoked magnetic field from ages 5 to 52 Donald C. Rojas*, Keeran Maharajh, Peter D. Teale, Michelle Ramos Kleman, Tara L. Benkers, Jon P. Carlson, Martin L. Reite Department of Psychiatry, University of Colorado Health Sciences Center, Box C268-68 CPH, 4200 E. 9th Avenue, Denver, CO 80262, USA Accepted 27 August 2005 Available online 28 November 2005
Abstract Objective: Adults exhibit strong auditory 40 Hz magnetic steady state responses (SSR). Although EEG measured SSR has been studied in children, the developmental course of the magnetic SSR is unknown. Methods: Sixty-nine healthy subjects ranging in age from 5 to 52 years participated in a magnetoencephalographic (MEG) study. Stimuli were monaural 500 ms duration click trains with a 25 ms inter-click interval. Contralateral magnetic responses for both hemispheres were recorded with a 37-channel MEG system. Responses were averaged and examined using wavelet-based time-frequency analysis. Source analyses were also conducted on a subset of the data. Results: Gamma power from 200 to 500 ms post-stimulus onset was computed and was significantly related to subject age in both hemispheres. Hemispheric asymmetry was observed for the anterior–posterior SSR source locations, suggestive of asymmetry similar to that previously described for the SSR and other auditory evoked magnetic field components. Conclusions: The 40 Hz power findings are generally consistent with previous EEG studies of steady state responses in children showing agerelated changes in the 40 Hz SSR. Significance: Age-related changes in the strength of the magnetic 40 Hz SSR may continue to develop well beyond early childhood, which should be taken into consideration in planning future studies using adolescents and young adults. q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Magnetoencephalography; Gamma band; Time-frequency analysis; Wavelets; Auditory evoked responses; Steady state evoked responses
1. Introduction Auditory steady state responses (SSR) can be elicited by a variety of periodic stimuli such as clicks or pure tone stimuli (Galambos et al., 1981; Hari et al., 1989; John et al., 2003; Maurizi et al., 1990). The SSR can also be produced by amplitude or frequency modulation (AM or FM) of continuous sounds (Boettcher et al., 2002; Picton et al., 1987; Rees et al., 1986; Ross et al., 2000). The auditory SSR to stimulus or modulation rates around 40 Hz, first described by Galambos et al. (1981) has been of particular interest because of its potential for assessing hearing thresholds (Galambos et al., 1981; Stapells et al., 1984). In adults, * Corresponding author. Tel.: C1 303 315 8624; fax: C1 303 315 5395. E-mail address: [email protected] (D.C. Rojas).
40 Hz stimulus or modulation rates generally produce the strongest SSR (Stapells et al., 1984). Although the 40 Hz SSR has been studied extensively in adults, very few data address the development of the response in childhood and adolescence. Although for adult subjects across a wide range of ages, stimulus or modulation rates of approximately 40/s generally produce the highest amplitude SSR (Boettcher et al., 2001; Boettcher et al., 2002; Hari et al., 1989) no such relationship has been observed in young children (Aoyagi et al., 1993; Shallop and Osterhammel, 1983; Stapells et al., 1988; Suzuki and Kobayashi, 1984). Suzuki et al. (1984), in a study of adults (ages 23–26 years), infants and children (ages 3 months–6 years) reported that, in contrast to adults, infants and children generally showed SSR amplitude decreases as stimulus rates increased beyond 20–30/s, a finding later
1388-2457/$30.00 q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2005.08.032
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
replicated by Stapells et al. (1988). Maurizi et al. (1990) recorded the 40 Hz SSR in 32 newborns and 10 children between the ages of 5 and 8 years, reporting that the SSR was obtained in 24 of 32 newborns examined at the strongest stimulus intensity studied. All 10 older children exhibited the SSR at the highest intensity level. In addition, test–retest reliability within a single recording session was examined, and the SSR reliability was reported to be unsatisfactory for the newborns, but better in the older children, leading the authors to conclude that 40 Hz SSR becomes more reliable and stable with increasing age. To our knowledge, there have not been any longitudinal developmental studies of the 40 Hz SSR, nor have there been any studies examining the response in adolescents. Although in adults, the 40 Hz SSR has been examined extensively using both electroencephalography (EEG) and magnetoencephalography (MEG), to date there have been no developmental studies of the response using MEG. This is potentially important because there is some evidence that the EEG recorded SSR is composed of different neural generators than the MEG recorded SSR (Schoonhoven et al., 2002) and preliminary, unpublished data from our laboratory suggested that there were significant developmental effects on the magnetic SSR to click stimuli presented at a rate of 40/s. We therefore studied children and adults in a cross-sectional examination of the effects of age on the magnetic 40 Hz SSR. We hypothesized that the strength of the magnetic 40 Hz SSR would increase with age, reaching an asymptotic level in early adulthood. We also hypothesized, based on our earlier studies of hemispheric asymmetry in auditory source localizations (Reite et al., 1995, 1997; Rojas et al., 1997; Teale et al., 2000, 2003), that the sources for the SSR would be more anterior in the right hemisphere than the left, and that this asymmetry would be greater in the male participants.
2. Methods We recorded steady state auditory evoked fields from 69 healthy individuals (32 female) ranging in age from 5 to 52 years (meanZ25.62, SDZ13.05 years, skewnessZ.32, kurtosisZK1.11). All subjects had normal hearing thresholds for the stimulus frequencies between 250 and 4000 Hz and were right-handed according to the Annett handedness criteria (Annett, 1985). None had significant neurological or psychiatric histories based on personal report. Participants gave their written informed consent for the procedures, as required by the Colorado Multiple Institutional Review Board, and were compensated for their time. Stimuli consisted of 2 ms duration bi-phasic pulses delivered every 25 ms for a total of 500 ms, as measured at the earpiece. These pulse trains were repeated 150 times every 1.5 s. Acoustic stimuli were produced using E.A.R. TONE 3A (Cabot Safety Corporation, Indianapolis IN)
111
transducers with 2 m of polyurethane tubing (3 mm inner diameter) and foam insert earpieces with 30 dB attenuation to exterior noise. Sound amplitude was adjusted to 65 dB SPL using a Bruel and Kjaer 2209 SPL meter and 4157 artificial ear. Magnetic field data were obtained with a Magnes I, 37 channel biomagnetometer (4-DNeuroImaging, San Diego, CA) using concentric rings of first-order axial gradiometers with 2 cm coil diameters and 5 cm baselines. The lower coils are arrayed on a spherical surface of 11 cm radius, and are spaced approximately 2 cm on center. Recordings were obtained from the contra-lateral hemisphere from the stimulated ear while the subjects were lying on their sides. Subjects watched a silent video of their own choice throughout the procedure to help them maintain a more consistent state of alertness. Data were collected using a 16 bit analog-to-digital converter with a sampling rate of 1041.7 Hz over a 1 s window with a 200 ms pre-stimulus period. Raw data were notch filtered at 60, 120 and 180 Hz using a custom built tracking (optically coupled phaselocked-loop clock derived from the AC mains), fourth order elliptic filter with 40 dB of attenuation. Additional bandpass filtering was achieved using a 1 Hz single pole analog high pass filter and a 200 Hz FIR low pass. Prior to recording, fiducial points (left and right preauriculars and the nasion) were digitized using the Polhemus 3SPACE tracker (Colchester, VT), and coil locations and orientations were then expressed in the coordinate system defined as: Y-axis along the line between the pre-auricular points, positive thru the left; X-axis perpendicular to the Y-axis at the midpoint and contained in the plane formed by the nasion and pre-auricular points; Z-axis perpendicular to the same plane, starting at the midpoint, and positive in the upward direction. The gradiometer array was positioned such that the center channel was over the zero-field crossing between the ingoing and outgoing magnetic extrema of the averaged auditory evoked field. This position was typically near an X-coordinate of 0 cm and near a Z-coordinate of 5 cm for the center channel of the array in the head-frame (i.e. approximately the C3 and C4 positions of the 10–20 International System). Following recording, the epoch files were edited visually, trial by trial, for eye blink, head movement and other artifacts, prior to averaging and all epochs with artifact contamination were excluded from further analyses. Fig. 1 (bottom row) illustrates the averaged waveforms for an adult and an adolescent that were bandpassed from 3 to 50 Hz using a phase invariant Butterworth filter. A continuous wavelet analysis utilizing the complex Morlet wavelet (wave number, 6) was then used to perform a time-frequency decomposition of the averaged data. Wavelet scales corresponding to frequencies from 1 to 90 Hz set at 1 Hz apart were used for the decomposition. There are different time and frequency uncertainties associated with various types of wavelets (Cohen, 1989).
112
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 1. Top row: wavelet-based time-frequency plots for an adult (A) and an adolescent (B) subject, ages 37 and 12 years, respectively. The color scaling in the plots indicates power relative to baseline for each corresponding frequency. Bottom row: overlapping, averaged auditory evoked field waveforms for each of the corresponding time-frequency plots in the top portion of the figure. Left hemisphere data are shown in both cases.
For the Morlet wavelet at the 40 Hz scale, one can expect a temporal spreading of power with a 10 dB maximum power reduction from a test impulse response (IR) signal of approximately 40 ms. To increase computational speed, a FFT was applied to the data and the Morlet wavelet whose product was then inverse Fourier transformed (Torrence and Compo, 1998). The data were baseline corrected at each frequency by the mean power calculated from a 45 to 145 ms pre-stimulus onset window. The channel with maximal power at 40 Hz was then chosen for subsequent analysis. The latency of maximal power from 200 to 500 ms and its full-width-half-maximum (FWHM) were then determined. This post-stimulus onset window was chosen to focus attention on the SSR rather than transient evoked components based on prior published data (Ross et al., 2002). Fig. 1 (top row) illustrates the results of the wavelet analysis after baseline correction. Source localization analysis was performed across the post-stimulus onset window from 10 to 600 ms using a single moving equivalent current dipole (ECD) in a conductive sphere model (Sarvas, 1987) and a sliding 5 ms wide time window (starting 2.5 ms before, and ending 2.5 ms after, each time point) at 1 ms increments for each hemisphere. The mean B-field amplitude for each channel for the 5 ms window was computed by averaging the samples in this interval. Custom, in-house software was employed to solve the inverse problem at each time point, as described in Teale et al. (2003). This procedure resulted in 590 ECD estimates per subject, per hemisphere. For each estimate, a goodness-of-fit parameter was computed as the mean squared error divided by the mean squared data and then subtracted from one (Kaukoranta et al., 1986). A search for minimum values in this data set from 50 to 550 ms typically yielded 40 time points—two for each stimulus pulse, separated by about 12.5 ms. It has been reported that these ECD estimates are consistent over the response
window (Hari et al., 1989; Pantev et al., 1994). Therefore, the dipole location coordinates were computed as means across these time points subject to the constraint that the fit be better than .80. This requirement reduced the number of time points to between 17 and 40 (left hemisphere mean: 33.72G4.74 dipoles; right hemisphere mean: 35.59G4.08 dipoles). Mean ECD estimates with Y-coordinate values less than 2 cm away from the midline in either hemisphere excluded a subject’s data from statistical analysis of the ECD locations. Datasets with fewer than 15 channels with a SNR of 1.2 or greater were excluded from the source analyses due to extremely poor goodness of fits obtained with low SNR data (i.e. near the biological noise floor). These criteria resulted in the exclusion of 29 data sets, resulting in a final N of 40 for the statistical analyses of the source parameters (Mean ageGSD: 28.99G12.54, range 7.57–51.44 years).
3. Results Statistical analyses were conducted using Statistica 6.0 (Statsoft, Tulsa, OK). All null hypothesis significance testing was two-tailed and conducted at .05 alpha. 3.1. 40 Hz SSR To examine the primary hypothesis concerning agerelated changes in SSR signal, separate regressions to a standard exponential rise to maximum equation were computed using the left and right 40 Hz relative power measures as dependent variables and age as the exponent term. Simple linear regressions were also computed to assess the adequacy of the exponential model compared to a simpler model of age-related change. The exponential regression for the left hemisphere was
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
significant, F(1, 68)Z14.43, P!.001, R2Z.18. The right hemisphere exponential regression was also significant, F(1, 68)Z7.34, P!.01, R2Z.10. The linear regressions were also significant [left: F(1,68)Z9.44, P!.01; right: F(1,68)Z7.84, P!.01], but explained less of the variance in the left hemisphere than the corresponding exponential fit (R2Z.12). The right hemisphere linear regression proportion of explained variance (R2Z.10) was comparable to the exponential fit result. For all four fits, the Kolmogorov–Smirnov test for normality of the distribution about the regression was passed. Fig. 2 illustrates the relationships between the left and right 40 Hz power measures and age. We also examined potential gender differences using independent Student’s t tests. These were computed separately for both left and right hemispheres and were both non-significant, t(67)Z1.25, PO.05 and t(67)Z1.07, PO.05, respectively.
113
To examine whether there was any evidence of SSR activity at other peak frequencies that might vary by age, we also examined the 200–500 ms period post-stimulus onset for peak frequency between 24 and 60 Hz. We chose these cutoff frequencies to exclude longer latency transient evoked responses (e.g. the M100) and because there was little evidence for any significant power above 60 Hz. There was no significant correlation between age and peak frequency for either left or right hemisphere (rZK.03, PO.05 and rZ.02, PO.05, respectively). Histograms of the peak frequencies identified show a normal distribution for the left hemisphere (Kolgomorov–Smirnov dZ.14, PO.10, meanZ39.80G6.91 Hz). For the right hemisphere, the distribution of peak frequency was not normal (dZ.19, P! .05, meanZ40.45G5.33 Hz), with a kurtosis of 4.19 indicating a much more peaked distribution than the distribution on the left, where the kurtosis was 1.59, indicating slightly less peakedness (Fig. 3). The clustering
Fig. 2. Scatterplot of relationships between left (top) and right (bottom) 40 Hz power (relative to baseline) and chronological age of participants. Exponential regression lines are also plotted as solid lines for both hemispheres.
114
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 3. Histogram of peak frequencies between 24 and 60 Hz identified from time-frequency analysis of 200–500 ms post-stimulus onset period, with fitted normal distributions for visual reference. Data from the left and right hemispheres are illustrated.
of values near 40 Hz is consistent with the driving of this frequency by the stimulus. 3.2. SSR source localization After eliminating SSR runs with low SNR from further analysis (see above), 40 subjects (20 female) remained for the source analyses. For these remaining subjects, the relationships between age and x, y and z dipole locations were examined in both hemispheres using Pearson r correlation coefficients. The correlation coefficients between age and left x, y and z were K.22, .12 and K.14, respectively (all PO.15). The corresponding right hemisphere coordinate correlations were .02, K.44 and K.23, of which only the y-coordinate correlation was significant, P! .005, and remained so even after Bonferonni correction for multiple comparisons. Given our hypothesis concerning anterior–posterior source asymmetry, we also correlated age with the subtraction of left from right x-coordinate. This correlation was also non-significant prior to multiple comparison correction (rZ.24, PO.10). To evaluate the source lateralization and gender difference hypotheses, we conducted a 2!2 mixed design ANOVA (gender by hemisphere), with hemisphere treated as a within-subjects variable and using the mean
x-coordinate (anterior–posterior location, described above) as the dependent measure. There was no significant gender main effect, F(1,38)Z2.05, PO.05. There was a significant hemisphere main effect, F(1,38)Z17.98, P!.0002, indicating that the right hemisphere source locations (.05G .81 cm) were located anteriorly to the left hemisphere locations (K.50G.66 cm) The gender by hemisphere interaction term was non-significant, F(1,38)Z.67, PO.05. For the y- and z-coordinates (medial–lateral and inferior– superior locations), separate 2!2 mixed model ANOVAs, identical to that used for the x-coordinate, were evaluated. For the y-coordinate analysis, since the sign of the coordinate indicates which hemisphere the source is located in, the absolute value of the y-coordinate was used in the ANOVA to avoid an artificially significant hemisphere effect due simply to the coordinate system convention employed. No significant main effects or interaction terms were noted for the y-coordinate. For the z-coordinate, the gender main effect was significant, F(1,38)Z5.61, P!.03, suggesting a slightly more inferiorly located source for the females (4.14G.90 cm) than the males (4.65G.83 cm). Neither the hemisphere main effect nor the interaction term, however, was significant for the z-coordinate. No other source parameter (orientation, strength or goodness of fit) was significantly different between genders or hemispheres.
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 4 illustrates the source analysis results from the left and right ear stimuli co-registered onto an MRI scan for that subject.
4. Discussion The majority of participants in this study exhibited a peak spectral power at 40 Hz between 200 and 500 ms poststimulus onset, relative to the pre-stimulus baseline, which is consistent with the driving frequency of the click-train stimulus. The tight distribution of peak frequencies around 40 Hz demonstrates the utility of wavelet-based timefrequency analyses applied to steady state evoked responses, particularly, in designing studies where one wants to examine whether the coupling between stimulus and cortex might be shifted (e.g. brain oscillates at slightly different frequency from stimulus). The data from the present study are consistent with our hypothesis that the amplitude of the magnetic 40 Hz SSR would increase with age. Our data provide some support for the hypothesized asymptotic SSR maximum in early adulthood, although the proportion of explained variance
115
in the exponential regression was only clearly superior to the linear regression in the left hemisphere. It seems unlikely, however, that the true relationship between age and 40 Hz SSR is strictly linear throughout the lifespan. More likely, our inability to establish a clear non-linear relationship between age and 40 Hz power in the right hemisphere reflects the lack of data from persons less than age five in the study, as well as higher variability in the SSR in adulthood. The increased variability of the SSR in adulthood might be attributable to unknown variability in the state of the participants, which was not strictly controlled. Previous EEG work on the 40 Hz SSR has shown that 40 Hz modulation of neurophysiological responses is particularly sensitive to the arousal state of the subject studied (Cohen et al., 1991; Picton et al., 2003; Plourde and Picton, 1990). We included the use of a participant selected video to encourage our subjects to maintain alertness for the recordings, but arousal and/or attention may have varied between subjects, and is an inherent limitation in passive stimulation studies such as this one. Although we believe that this is the first study to examine the 40 Hz SSR in children and adolescents using MEG,
Fig. 4. Overlay of source analyses of data for subject illustrated in Fig. 1(A) onto co-registered MRI scan.
116
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
these data are essentially in agreement with previous EEG studies on children and adults, which together suggest that the 40 Hz SSR is sensitive to maturational effects. The picture is less clear, however, when the question of timing of those changes is asked. A single EEG study examining the SSR in newborns and children ages 5–8 reported prominent 40 Hz SSR in all of the older children (Maurizi et al., 1990), but the small sample size and lack of inclusion of an adult comparison group prevent estimation of age of maturity from the data. Nonetheless, the magnetic responses of the younger children from the current study, whose ages overlap with the Maurizi et al. (1990) sample, certainly were not robust in comparison. There are several possibilities for the discrepancy between those estimates and the current study, which suggests further maturation into adulthood. First, there is emerging evidence that the EEG and MEG 40 Hz SSR differently reflect the contribution of several generators (Schoonhoven et al., 2002). This is not surprising considering the general differences between EEG and MEG with respect to the measured signal (e.g. radially versus tangentially oriented sources, intraneuronal versus volume conducted currents, etc.). In addition, steady state stimuli are known to activate multiple regions of the auditory pathway (Joris et al., 2004). Different regions of the auditory system appear to have best modulation frequencies, and in general, the higher the modulation frequency the lower the level of the auditory system that responds (Giraud et al., 2000; Harms and Melcher, 2002; Herdman et al., 2002; Joris et al., 2004). For the 40 Hz SSR, it has been suggested on the basis of EEG source analyses that contributions are seen primarily from overlapping responses from the auditory thalamus and auditory cortex (Herdman et al., 2002), although MEG source analyses consistently exhibit bilateral auditory cortical sources (Gutschalk et al., 1999; Hari et al., 1989; Pantev et al., 1993, 1996), possibly due to the relative insensitivity of the MEG measurements to deeper sources. It should be noted that one of the limitations of this study is the use of a small detector array, which precludes serious investigation of either multiple source models or distributed source modeling approaches. The EEG literature, taken together with the current MEG study, suggests the possibility of differential maturation of regions of the auditory system responsible for generating the 40 Hz SSR and that EEG and MEG measurements may be preferentially sensitive to the development of those generators. Simultaneous measurement of the auditory evoked magnetic and electric SSR in children will be necessary to test this hypothesis, which is highly speculative at this point. One prediction emerging from this comparison would be that EEG 40 Hz responses would mature earlier, based both on the empirical findings to date and the plausible idea that EEG 40 Hz sources are reflecting neuronal generators that are known to mature earlier (e.g. brain-stem and thalamus) than the MEG generators (e.g. cerebral cortex).
The anterior–posterior asymmetry of the SSR generators is consistent with prior literature on both the 40 Hz SSR (Teale et al., 2003) and the lower frequency auditory evoked responses such as the M50 and M100 (Reite et al., 1988, 1989; Teale et al., 1998). We observed no relationship between this asymmetry and age in our sample, which might be consistent with the findings of asymmetry of the underlying anatomical regions at birth (Witelson and Pallie, 1973). The precise nature of the relationship of the source generators of the magnetic SSR to anatomical structures is unknown, however, and this would need to be explored directly in future studies. There are several other limitations to the current study. First, we did not include children younger than age five in our sample. This limits our ability to directly compare our findings to the published developmental literature on 40 Hz SSR, which largely focuses on the period between birth and 5–8 years of age. Second, we did not employ multiple stimulation rates in this study. There is evidence from EEG SSR data that both lower and higher modulation frequencies than 40 Hz produce more robust responses from children (Stapells et al., 1988; Suzuki and Kobayashi, 1984). Finally, as previously discussed, the lack of stricter control over the state of arousal and/or attention in our participants may have increased the variability in our data, given the statedependency of the 40 Hz SSR. Future work is planned to address these shortcomings.
Acknowledgements Support for this project was provided by Public Health Service grants MH60214, MH47476, MH63442 and the National Alliance for Autism Research.
References Annett M. Left, right, hand and brain: the right shift theory. Hillsdale, NJ: Lawrence Erlbaum; 1985. Aoyagi M, Kiren T, Kim Y, Suzuki Y, Fuse T, Koike Y. Optimal modulation frequency for amplitude-modulation following response in young children during sleep. Hear Res 1993;65(1–2):253–61. Boettcher FA, Poth EA, Mills JH, Dubno JR. The amplitude-modulation following response in young and aged human subjects. Hear Res 2001; 153(1–2):32–42. Boettcher FA, Madhotra D, Poth EA, Mills JH. The frequency-modulation following response in young and aged human subjects. Hear Res 2002; 165(1–2):10–18. Cohen L. Time frequency-distributions—a review. Proc IEEE 1989;77: 941–81. Cohen LT, Rickards FW, Clark GM. A comparison of steady-state evoked potentials to modulated tones in awake and sleeping humans. J Acoust Soc Am 1991;90(5):2467–79. Galambos R, Makeig S, Talmachoff PJ. A 40-hz auditory potential recorded from the human scalp. Proc Natl Acad Sci USA 1981;78(4):2643–7. Giraud AL, Lorenzi C, Ashburner J, Wable J, Johnsrude I, Frackowiak R, Kleinschmidt A. Representation of the temporal envelope of sounds in the human brain. J Neurophysiol 2000;84(3):1588–98.
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117 Gutschalk A, Mase R, Roth R, Ille N, Rupp A, Hahnel S, Picton TW, Scherg M. Deconvolution of 40 Hz steady-state fields reveals two overlapping source activities of the human auditory cortex. Clin Neurophysiol 1999;110(5):856–68. Hari R, Hamalainen M, Joutsiniemi SL. Neuromagnetic steady-state responses to auditory stimuli. J Acoust Soc Am 1989;86(3):1033–9. Harms MP, Melcher JR. Sound repetition rate in the human auditory pathway: representations in the waveshape and amplitude of fMRI activation. J Neurophysiol 2002;88(3):1433–50. Herdman AT, Lins O, Van Roon P, Stapells DR, Scherg M, Picton TW. Intracerebral sources of human auditory steady-state responses. Brain Topogr 2002;15(2):69–86. John MS, Dimitrijevic A, Picton TW. Efficient stimuli for evoking auditory steady-state responses. Ear Hear 2003;24(5):406–23. Joris PX, Schreiner CE, Rees A. Neural processing of amplitude-modulated sounds. Physiol Rev 2004;84(2):541–77. Kaukoranta E, Hamalainen M, Sarvas J, Hari R. Mixed and sensory nerve stimulations activate different cytoarchitectonic areas in the human primary somatosensory cortex SI. Neuromagnetic recordings and statistical considerations. Exp Brain Res 1986;63(1):60–6. Maurizi M, Almadori G, Paludetti G, Ottaviani F, Rosignoli M, Luciano R. 40-Hz steady-state responses in newborns and in children. Audiology 1990;29(6):322–8. Pantev C, Elbert T, Makeig S, Hampson S, Eulitz C, Hoke M. Relationship of transient and steady-state auditory evoked fields. Electroencephalogr Clin Neurophysiol 1993;88(5):389–96. Pantev C, Eulitz C, Elbert T, Hoke M. The auditory evoked sustained field: origin and frequency dependence. Electroencephalogr Clin Neurophysiol 1994;90(1):82–90. Pantev C, Roberts LE, Elbert T, Ross B, Wienbruch C. Tonotopic organization of the sources of human auditory steady-state responses. Hear Res 1996;101(1–2):62–74. Picton TW, Skinner CR, Champagne SC, Kellett AJ, Maiste AC. Potentials evoked by the sinusoidal modulation of the amplitude or frequency of a tone. J Acoust Soc Am 1987;82(1):165–78. Picton TW, John MS, Purcell DW, Plourde G. Human auditory steady-state responses: the effects of recording technique and state of arousal. Anesth Analg 2003;97(5):1396–402. Plourde G, Picton TW. Human auditory steady-state response during general anesthesia. Anesth Analg 1990;71(5):460–8. Rees A, Green GG, Kay RH. Steady-state evoked responses to sinusoidally amplitude-modulated sounds recorded in man. Hear Res 1986;23(2): 123–33. Reite M, Teale P, Zimmerman J, Davis K, Whalen J. Source location of a 50 msec latency auditory evoked field component. Electroencephalogr Clin Neurophysiol 1988;70(6):490–8. Reite M, Teale P, Goldstein L, Whalen J, Linnville S. Late auditory magnetic sources may differ in the left hemisphere of schizophrenic patients. Arch Gen Psychiatry 1989;46:565–72.
117
Reite M, Sheeder J, Teale P, Richardson D, Adams M, Simon J. Meg based brain laterality: sex differences in normal adults. Neuropsychologia 1995;33(12):1607–16. Reite M, Sheeder J, Teale P, Adams M, Richardson D, Simon J, Jones RH, Rojas DC. Magnetic source imaging evidence of sex differences in cerebral lateralization in schizophrenia. Arch Gen Psychiatry 1997;54: 433–40. Rojas DC, Teale P, Sheeder J, Simon J, Reite M. Sex-specific expression of heschl’s gyrus functional and structural abnormalities in paranoid schizophrenia. Am J Psychiatry 1997;154(12):1655–62. Ross B, Borgmann C, Draganova R, Roberts LE, Pantev C. A highprecision magnetoencephalographic study of human auditory steadystate responses to amplitude-modulated tones. J Acoust Soc Am 2000; 108(2):679–91. Ross B, Picton TW, Pantev C. Temporal integration in the human auditory cortex as represented by the development of the steady-state magnetic field. Hear Res 2002;165(1–2):68–84. Sarvas J. Basic mathematical and electromagnetic concepts of the biomagnetic inverse problem. Phys Med Biol 1987;1:11–22. Schoonhoven R, Boden CJR, Verbunt JPA, De Munck JC. Sources of meg and eeg auditory steady-state responses. In: Nowak H, Haueisen J, Geissler F, Huonker R, editors. Biomag 2002: proceedings of the 13th international conference on biomagnetism. Berlin: VDE Verlag; 2002. p. 116–8. Shallop JK, Osterhammel PA. A comparative study of measurements of SN-10 and the 40/sec middle latency responses in newborns. Scand Audiol 1983;12(2):91–5. Stapells DR, Linden D, Suffield JB, Hamel G, Picton TW. Human auditory steady state potentials. Ear Hear 1984;5(2):105–13. Stapells DR, Galambos R, Costello JA, Makeig S. Inconsistency of auditory middle latency and steady-state responses in infants. Electroencephalogr Clin Neurophysiol 1988;71(4):289–95. Suzuki T, Kobayashi K. An evaluation of 40-Hz event-related potentials in young children. Audiology 1984;23(6):599–604. Teale P, Sheeder J, Rojas DC, Walker J, Reite M. Sequential source model of the m100 exhibits inter-hemispheric asymmetry. NeuroReport 1998; 9:2647–52. Teale P, Reite M, Rojas DC, Sheeder J, Arciniegas D. Fine structure of the auditory m100 in schizophrenia and schizoaffective disorder. Biol Psychiatry 2000;48(11):1109–12. Teale P, Carlson J, Rojas D, Reite M. Reduced laterality of the source locations for generators of the auditory steady-state field in schizophrenia. Biol Psychiatry 2003;54(11):1149–53. Torrence C, Compo GP. A practical guide to wavelet analysis. Bull Am Meteorol Soc 1998;79:61–78. Witelson SF, Pallie W. Left hemisphere specialization for language in the newborn: neuroanatomical evidence of asymmetry. Brain 1973;96: 641–6.
Development of the 40 Hz steady state auditory evoked magnetic field from ages 5 to 52 Donald C. Rojas*, Keeran Maharajh, Peter D. Teale, Michelle Ramos Kleman, Tara L. Benkers, Jon P. Carlson, Martin L. Reite Department of Psychiatry, University of Colorado Health Sciences Center, Box C268-68 CPH, 4200 E. 9th Avenue, Denver, CO 80262, USA Accepted 27 August 2005 Available online 28 November 2005
Abstract Objective: Adults exhibit strong auditory 40 Hz magnetic steady state responses (SSR). Although EEG measured SSR has been studied in children, the developmental course of the magnetic SSR is unknown. Methods: Sixty-nine healthy subjects ranging in age from 5 to 52 years participated in a magnetoencephalographic (MEG) study. Stimuli were monaural 500 ms duration click trains with a 25 ms inter-click interval. Contralateral magnetic responses for both hemispheres were recorded with a 37-channel MEG system. Responses were averaged and examined using wavelet-based time-frequency analysis. Source analyses were also conducted on a subset of the data. Results: Gamma power from 200 to 500 ms post-stimulus onset was computed and was significantly related to subject age in both hemispheres. Hemispheric asymmetry was observed for the anterior–posterior SSR source locations, suggestive of asymmetry similar to that previously described for the SSR and other auditory evoked magnetic field components. Conclusions: The 40 Hz power findings are generally consistent with previous EEG studies of steady state responses in children showing agerelated changes in the 40 Hz SSR. Significance: Age-related changes in the strength of the magnetic 40 Hz SSR may continue to develop well beyond early childhood, which should be taken into consideration in planning future studies using adolescents and young adults. q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Magnetoencephalography; Gamma band; Time-frequency analysis; Wavelets; Auditory evoked responses; Steady state evoked responses
1. Introduction Auditory steady state responses (SSR) can be elicited by a variety of periodic stimuli such as clicks or pure tone stimuli (Galambos et al., 1981; Hari et al., 1989; John et al., 2003; Maurizi et al., 1990). The SSR can also be produced by amplitude or frequency modulation (AM or FM) of continuous sounds (Boettcher et al., 2002; Picton et al., 1987; Rees et al., 1986; Ross et al., 2000). The auditory SSR to stimulus or modulation rates around 40 Hz, first described by Galambos et al. (1981) has been of particular interest because of its potential for assessing hearing thresholds (Galambos et al., 1981; Stapells et al., 1984). In adults, * Corresponding author. Tel.: C1 303 315 8624; fax: C1 303 315 5395. E-mail address: [email protected] (D.C. Rojas).
40 Hz stimulus or modulation rates generally produce the strongest SSR (Stapells et al., 1984). Although the 40 Hz SSR has been studied extensively in adults, very few data address the development of the response in childhood and adolescence. Although for adult subjects across a wide range of ages, stimulus or modulation rates of approximately 40/s generally produce the highest amplitude SSR (Boettcher et al., 2001; Boettcher et al., 2002; Hari et al., 1989) no such relationship has been observed in young children (Aoyagi et al., 1993; Shallop and Osterhammel, 1983; Stapells et al., 1988; Suzuki and Kobayashi, 1984). Suzuki et al. (1984), in a study of adults (ages 23–26 years), infants and children (ages 3 months–6 years) reported that, in contrast to adults, infants and children generally showed SSR amplitude decreases as stimulus rates increased beyond 20–30/s, a finding later
1388-2457/$30.00 q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2005.08.032
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
replicated by Stapells et al. (1988). Maurizi et al. (1990) recorded the 40 Hz SSR in 32 newborns and 10 children between the ages of 5 and 8 years, reporting that the SSR was obtained in 24 of 32 newborns examined at the strongest stimulus intensity studied. All 10 older children exhibited the SSR at the highest intensity level. In addition, test–retest reliability within a single recording session was examined, and the SSR reliability was reported to be unsatisfactory for the newborns, but better in the older children, leading the authors to conclude that 40 Hz SSR becomes more reliable and stable with increasing age. To our knowledge, there have not been any longitudinal developmental studies of the 40 Hz SSR, nor have there been any studies examining the response in adolescents. Although in adults, the 40 Hz SSR has been examined extensively using both electroencephalography (EEG) and magnetoencephalography (MEG), to date there have been no developmental studies of the response using MEG. This is potentially important because there is some evidence that the EEG recorded SSR is composed of different neural generators than the MEG recorded SSR (Schoonhoven et al., 2002) and preliminary, unpublished data from our laboratory suggested that there were significant developmental effects on the magnetic SSR to click stimuli presented at a rate of 40/s. We therefore studied children and adults in a cross-sectional examination of the effects of age on the magnetic 40 Hz SSR. We hypothesized that the strength of the magnetic 40 Hz SSR would increase with age, reaching an asymptotic level in early adulthood. We also hypothesized, based on our earlier studies of hemispheric asymmetry in auditory source localizations (Reite et al., 1995, 1997; Rojas et al., 1997; Teale et al., 2000, 2003), that the sources for the SSR would be more anterior in the right hemisphere than the left, and that this asymmetry would be greater in the male participants.
2. Methods We recorded steady state auditory evoked fields from 69 healthy individuals (32 female) ranging in age from 5 to 52 years (meanZ25.62, SDZ13.05 years, skewnessZ.32, kurtosisZK1.11). All subjects had normal hearing thresholds for the stimulus frequencies between 250 and 4000 Hz and were right-handed according to the Annett handedness criteria (Annett, 1985). None had significant neurological or psychiatric histories based on personal report. Participants gave their written informed consent for the procedures, as required by the Colorado Multiple Institutional Review Board, and were compensated for their time. Stimuli consisted of 2 ms duration bi-phasic pulses delivered every 25 ms for a total of 500 ms, as measured at the earpiece. These pulse trains were repeated 150 times every 1.5 s. Acoustic stimuli were produced using E.A.R. TONE 3A (Cabot Safety Corporation, Indianapolis IN)
111
transducers with 2 m of polyurethane tubing (3 mm inner diameter) and foam insert earpieces with 30 dB attenuation to exterior noise. Sound amplitude was adjusted to 65 dB SPL using a Bruel and Kjaer 2209 SPL meter and 4157 artificial ear. Magnetic field data were obtained with a Magnes I, 37 channel biomagnetometer (4-DNeuroImaging, San Diego, CA) using concentric rings of first-order axial gradiometers with 2 cm coil diameters and 5 cm baselines. The lower coils are arrayed on a spherical surface of 11 cm radius, and are spaced approximately 2 cm on center. Recordings were obtained from the contra-lateral hemisphere from the stimulated ear while the subjects were lying on their sides. Subjects watched a silent video of their own choice throughout the procedure to help them maintain a more consistent state of alertness. Data were collected using a 16 bit analog-to-digital converter with a sampling rate of 1041.7 Hz over a 1 s window with a 200 ms pre-stimulus period. Raw data were notch filtered at 60, 120 and 180 Hz using a custom built tracking (optically coupled phaselocked-loop clock derived from the AC mains), fourth order elliptic filter with 40 dB of attenuation. Additional bandpass filtering was achieved using a 1 Hz single pole analog high pass filter and a 200 Hz FIR low pass. Prior to recording, fiducial points (left and right preauriculars and the nasion) were digitized using the Polhemus 3SPACE tracker (Colchester, VT), and coil locations and orientations were then expressed in the coordinate system defined as: Y-axis along the line between the pre-auricular points, positive thru the left; X-axis perpendicular to the Y-axis at the midpoint and contained in the plane formed by the nasion and pre-auricular points; Z-axis perpendicular to the same plane, starting at the midpoint, and positive in the upward direction. The gradiometer array was positioned such that the center channel was over the zero-field crossing between the ingoing and outgoing magnetic extrema of the averaged auditory evoked field. This position was typically near an X-coordinate of 0 cm and near a Z-coordinate of 5 cm for the center channel of the array in the head-frame (i.e. approximately the C3 and C4 positions of the 10–20 International System). Following recording, the epoch files were edited visually, trial by trial, for eye blink, head movement and other artifacts, prior to averaging and all epochs with artifact contamination were excluded from further analyses. Fig. 1 (bottom row) illustrates the averaged waveforms for an adult and an adolescent that were bandpassed from 3 to 50 Hz using a phase invariant Butterworth filter. A continuous wavelet analysis utilizing the complex Morlet wavelet (wave number, 6) was then used to perform a time-frequency decomposition of the averaged data. Wavelet scales corresponding to frequencies from 1 to 90 Hz set at 1 Hz apart were used for the decomposition. There are different time and frequency uncertainties associated with various types of wavelets (Cohen, 1989).
112
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 1. Top row: wavelet-based time-frequency plots for an adult (A) and an adolescent (B) subject, ages 37 and 12 years, respectively. The color scaling in the plots indicates power relative to baseline for each corresponding frequency. Bottom row: overlapping, averaged auditory evoked field waveforms for each of the corresponding time-frequency plots in the top portion of the figure. Left hemisphere data are shown in both cases.
For the Morlet wavelet at the 40 Hz scale, one can expect a temporal spreading of power with a 10 dB maximum power reduction from a test impulse response (IR) signal of approximately 40 ms. To increase computational speed, a FFT was applied to the data and the Morlet wavelet whose product was then inverse Fourier transformed (Torrence and Compo, 1998). The data were baseline corrected at each frequency by the mean power calculated from a 45 to 145 ms pre-stimulus onset window. The channel with maximal power at 40 Hz was then chosen for subsequent analysis. The latency of maximal power from 200 to 500 ms and its full-width-half-maximum (FWHM) were then determined. This post-stimulus onset window was chosen to focus attention on the SSR rather than transient evoked components based on prior published data (Ross et al., 2002). Fig. 1 (top row) illustrates the results of the wavelet analysis after baseline correction. Source localization analysis was performed across the post-stimulus onset window from 10 to 600 ms using a single moving equivalent current dipole (ECD) in a conductive sphere model (Sarvas, 1987) and a sliding 5 ms wide time window (starting 2.5 ms before, and ending 2.5 ms after, each time point) at 1 ms increments for each hemisphere. The mean B-field amplitude for each channel for the 5 ms window was computed by averaging the samples in this interval. Custom, in-house software was employed to solve the inverse problem at each time point, as described in Teale et al. (2003). This procedure resulted in 590 ECD estimates per subject, per hemisphere. For each estimate, a goodness-of-fit parameter was computed as the mean squared error divided by the mean squared data and then subtracted from one (Kaukoranta et al., 1986). A search for minimum values in this data set from 50 to 550 ms typically yielded 40 time points—two for each stimulus pulse, separated by about 12.5 ms. It has been reported that these ECD estimates are consistent over the response
window (Hari et al., 1989; Pantev et al., 1994). Therefore, the dipole location coordinates were computed as means across these time points subject to the constraint that the fit be better than .80. This requirement reduced the number of time points to between 17 and 40 (left hemisphere mean: 33.72G4.74 dipoles; right hemisphere mean: 35.59G4.08 dipoles). Mean ECD estimates with Y-coordinate values less than 2 cm away from the midline in either hemisphere excluded a subject’s data from statistical analysis of the ECD locations. Datasets with fewer than 15 channels with a SNR of 1.2 or greater were excluded from the source analyses due to extremely poor goodness of fits obtained with low SNR data (i.e. near the biological noise floor). These criteria resulted in the exclusion of 29 data sets, resulting in a final N of 40 for the statistical analyses of the source parameters (Mean ageGSD: 28.99G12.54, range 7.57–51.44 years).
3. Results Statistical analyses were conducted using Statistica 6.0 (Statsoft, Tulsa, OK). All null hypothesis significance testing was two-tailed and conducted at .05 alpha. 3.1. 40 Hz SSR To examine the primary hypothesis concerning agerelated changes in SSR signal, separate regressions to a standard exponential rise to maximum equation were computed using the left and right 40 Hz relative power measures as dependent variables and age as the exponent term. Simple linear regressions were also computed to assess the adequacy of the exponential model compared to a simpler model of age-related change. The exponential regression for the left hemisphere was
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
significant, F(1, 68)Z14.43, P!.001, R2Z.18. The right hemisphere exponential regression was also significant, F(1, 68)Z7.34, P!.01, R2Z.10. The linear regressions were also significant [left: F(1,68)Z9.44, P!.01; right: F(1,68)Z7.84, P!.01], but explained less of the variance in the left hemisphere than the corresponding exponential fit (R2Z.12). The right hemisphere linear regression proportion of explained variance (R2Z.10) was comparable to the exponential fit result. For all four fits, the Kolmogorov–Smirnov test for normality of the distribution about the regression was passed. Fig. 2 illustrates the relationships between the left and right 40 Hz power measures and age. We also examined potential gender differences using independent Student’s t tests. These were computed separately for both left and right hemispheres and were both non-significant, t(67)Z1.25, PO.05 and t(67)Z1.07, PO.05, respectively.
113
To examine whether there was any evidence of SSR activity at other peak frequencies that might vary by age, we also examined the 200–500 ms period post-stimulus onset for peak frequency between 24 and 60 Hz. We chose these cutoff frequencies to exclude longer latency transient evoked responses (e.g. the M100) and because there was little evidence for any significant power above 60 Hz. There was no significant correlation between age and peak frequency for either left or right hemisphere (rZK.03, PO.05 and rZ.02, PO.05, respectively). Histograms of the peak frequencies identified show a normal distribution for the left hemisphere (Kolgomorov–Smirnov dZ.14, PO.10, meanZ39.80G6.91 Hz). For the right hemisphere, the distribution of peak frequency was not normal (dZ.19, P! .05, meanZ40.45G5.33 Hz), with a kurtosis of 4.19 indicating a much more peaked distribution than the distribution on the left, where the kurtosis was 1.59, indicating slightly less peakedness (Fig. 3). The clustering
Fig. 2. Scatterplot of relationships between left (top) and right (bottom) 40 Hz power (relative to baseline) and chronological age of participants. Exponential regression lines are also plotted as solid lines for both hemispheres.
114
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 3. Histogram of peak frequencies between 24 and 60 Hz identified from time-frequency analysis of 200–500 ms post-stimulus onset period, with fitted normal distributions for visual reference. Data from the left and right hemispheres are illustrated.
of values near 40 Hz is consistent with the driving of this frequency by the stimulus. 3.2. SSR source localization After eliminating SSR runs with low SNR from further analysis (see above), 40 subjects (20 female) remained for the source analyses. For these remaining subjects, the relationships between age and x, y and z dipole locations were examined in both hemispheres using Pearson r correlation coefficients. The correlation coefficients between age and left x, y and z were K.22, .12 and K.14, respectively (all PO.15). The corresponding right hemisphere coordinate correlations were .02, K.44 and K.23, of which only the y-coordinate correlation was significant, P! .005, and remained so even after Bonferonni correction for multiple comparisons. Given our hypothesis concerning anterior–posterior source asymmetry, we also correlated age with the subtraction of left from right x-coordinate. This correlation was also non-significant prior to multiple comparison correction (rZ.24, PO.10). To evaluate the source lateralization and gender difference hypotheses, we conducted a 2!2 mixed design ANOVA (gender by hemisphere), with hemisphere treated as a within-subjects variable and using the mean
x-coordinate (anterior–posterior location, described above) as the dependent measure. There was no significant gender main effect, F(1,38)Z2.05, PO.05. There was a significant hemisphere main effect, F(1,38)Z17.98, P!.0002, indicating that the right hemisphere source locations (.05G .81 cm) were located anteriorly to the left hemisphere locations (K.50G.66 cm) The gender by hemisphere interaction term was non-significant, F(1,38)Z.67, PO.05. For the y- and z-coordinates (medial–lateral and inferior– superior locations), separate 2!2 mixed model ANOVAs, identical to that used for the x-coordinate, were evaluated. For the y-coordinate analysis, since the sign of the coordinate indicates which hemisphere the source is located in, the absolute value of the y-coordinate was used in the ANOVA to avoid an artificially significant hemisphere effect due simply to the coordinate system convention employed. No significant main effects or interaction terms were noted for the y-coordinate. For the z-coordinate, the gender main effect was significant, F(1,38)Z5.61, P!.03, suggesting a slightly more inferiorly located source for the females (4.14G.90 cm) than the males (4.65G.83 cm). Neither the hemisphere main effect nor the interaction term, however, was significant for the z-coordinate. No other source parameter (orientation, strength or goodness of fit) was significantly different between genders or hemispheres.
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
Fig. 4 illustrates the source analysis results from the left and right ear stimuli co-registered onto an MRI scan for that subject.
4. Discussion The majority of participants in this study exhibited a peak spectral power at 40 Hz between 200 and 500 ms poststimulus onset, relative to the pre-stimulus baseline, which is consistent with the driving frequency of the click-train stimulus. The tight distribution of peak frequencies around 40 Hz demonstrates the utility of wavelet-based timefrequency analyses applied to steady state evoked responses, particularly, in designing studies where one wants to examine whether the coupling between stimulus and cortex might be shifted (e.g. brain oscillates at slightly different frequency from stimulus). The data from the present study are consistent with our hypothesis that the amplitude of the magnetic 40 Hz SSR would increase with age. Our data provide some support for the hypothesized asymptotic SSR maximum in early adulthood, although the proportion of explained variance
115
in the exponential regression was only clearly superior to the linear regression in the left hemisphere. It seems unlikely, however, that the true relationship between age and 40 Hz SSR is strictly linear throughout the lifespan. More likely, our inability to establish a clear non-linear relationship between age and 40 Hz power in the right hemisphere reflects the lack of data from persons less than age five in the study, as well as higher variability in the SSR in adulthood. The increased variability of the SSR in adulthood might be attributable to unknown variability in the state of the participants, which was not strictly controlled. Previous EEG work on the 40 Hz SSR has shown that 40 Hz modulation of neurophysiological responses is particularly sensitive to the arousal state of the subject studied (Cohen et al., 1991; Picton et al., 2003; Plourde and Picton, 1990). We included the use of a participant selected video to encourage our subjects to maintain alertness for the recordings, but arousal and/or attention may have varied between subjects, and is an inherent limitation in passive stimulation studies such as this one. Although we believe that this is the first study to examine the 40 Hz SSR in children and adolescents using MEG,
Fig. 4. Overlay of source analyses of data for subject illustrated in Fig. 1(A) onto co-registered MRI scan.
116
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117
these data are essentially in agreement with previous EEG studies on children and adults, which together suggest that the 40 Hz SSR is sensitive to maturational effects. The picture is less clear, however, when the question of timing of those changes is asked. A single EEG study examining the SSR in newborns and children ages 5–8 reported prominent 40 Hz SSR in all of the older children (Maurizi et al., 1990), but the small sample size and lack of inclusion of an adult comparison group prevent estimation of age of maturity from the data. Nonetheless, the magnetic responses of the younger children from the current study, whose ages overlap with the Maurizi et al. (1990) sample, certainly were not robust in comparison. There are several possibilities for the discrepancy between those estimates and the current study, which suggests further maturation into adulthood. First, there is emerging evidence that the EEG and MEG 40 Hz SSR differently reflect the contribution of several generators (Schoonhoven et al., 2002). This is not surprising considering the general differences between EEG and MEG with respect to the measured signal (e.g. radially versus tangentially oriented sources, intraneuronal versus volume conducted currents, etc.). In addition, steady state stimuli are known to activate multiple regions of the auditory pathway (Joris et al., 2004). Different regions of the auditory system appear to have best modulation frequencies, and in general, the higher the modulation frequency the lower the level of the auditory system that responds (Giraud et al., 2000; Harms and Melcher, 2002; Herdman et al., 2002; Joris et al., 2004). For the 40 Hz SSR, it has been suggested on the basis of EEG source analyses that contributions are seen primarily from overlapping responses from the auditory thalamus and auditory cortex (Herdman et al., 2002), although MEG source analyses consistently exhibit bilateral auditory cortical sources (Gutschalk et al., 1999; Hari et al., 1989; Pantev et al., 1993, 1996), possibly due to the relative insensitivity of the MEG measurements to deeper sources. It should be noted that one of the limitations of this study is the use of a small detector array, which precludes serious investigation of either multiple source models or distributed source modeling approaches. The EEG literature, taken together with the current MEG study, suggests the possibility of differential maturation of regions of the auditory system responsible for generating the 40 Hz SSR and that EEG and MEG measurements may be preferentially sensitive to the development of those generators. Simultaneous measurement of the auditory evoked magnetic and electric SSR in children will be necessary to test this hypothesis, which is highly speculative at this point. One prediction emerging from this comparison would be that EEG 40 Hz responses would mature earlier, based both on the empirical findings to date and the plausible idea that EEG 40 Hz sources are reflecting neuronal generators that are known to mature earlier (e.g. brain-stem and thalamus) than the MEG generators (e.g. cerebral cortex).
The anterior–posterior asymmetry of the SSR generators is consistent with prior literature on both the 40 Hz SSR (Teale et al., 2003) and the lower frequency auditory evoked responses such as the M50 and M100 (Reite et al., 1988, 1989; Teale et al., 1998). We observed no relationship between this asymmetry and age in our sample, which might be consistent with the findings of asymmetry of the underlying anatomical regions at birth (Witelson and Pallie, 1973). The precise nature of the relationship of the source generators of the magnetic SSR to anatomical structures is unknown, however, and this would need to be explored directly in future studies. There are several other limitations to the current study. First, we did not include children younger than age five in our sample. This limits our ability to directly compare our findings to the published developmental literature on 40 Hz SSR, which largely focuses on the period between birth and 5–8 years of age. Second, we did not employ multiple stimulation rates in this study. There is evidence from EEG SSR data that both lower and higher modulation frequencies than 40 Hz produce more robust responses from children (Stapells et al., 1988; Suzuki and Kobayashi, 1984). Finally, as previously discussed, the lack of stricter control over the state of arousal and/or attention in our participants may have increased the variability in our data, given the statedependency of the 40 Hz SSR. Future work is planned to address these shortcomings.
Acknowledgements Support for this project was provided by Public Health Service grants MH60214, MH47476, MH63442 and the National Alliance for Autism Research.
References Annett M. Left, right, hand and brain: the right shift theory. Hillsdale, NJ: Lawrence Erlbaum; 1985. Aoyagi M, Kiren T, Kim Y, Suzuki Y, Fuse T, Koike Y. Optimal modulation frequency for amplitude-modulation following response in young children during sleep. Hear Res 1993;65(1–2):253–61. Boettcher FA, Poth EA, Mills JH, Dubno JR. The amplitude-modulation following response in young and aged human subjects. Hear Res 2001; 153(1–2):32–42. Boettcher FA, Madhotra D, Poth EA, Mills JH. The frequency-modulation following response in young and aged human subjects. Hear Res 2002; 165(1–2):10–18. Cohen L. Time frequency-distributions—a review. Proc IEEE 1989;77: 941–81. Cohen LT, Rickards FW, Clark GM. A comparison of steady-state evoked potentials to modulated tones in awake and sleeping humans. J Acoust Soc Am 1991;90(5):2467–79. Galambos R, Makeig S, Talmachoff PJ. A 40-hz auditory potential recorded from the human scalp. Proc Natl Acad Sci USA 1981;78(4):2643–7. Giraud AL, Lorenzi C, Ashburner J, Wable J, Johnsrude I, Frackowiak R, Kleinschmidt A. Representation of the temporal envelope of sounds in the human brain. J Neurophysiol 2000;84(3):1588–98.
D.C. Rojas et al. / Clinical Neurophysiology 117 (2006) 110–117 Gutschalk A, Mase R, Roth R, Ille N, Rupp A, Hahnel S, Picton TW, Scherg M. Deconvolution of 40 Hz steady-state fields reveals two overlapping source activities of the human auditory cortex. Clin Neurophysiol 1999;110(5):856–68. Hari R, Hamalainen M, Joutsiniemi SL. Neuromagnetic steady-state responses to auditory stimuli. J Acoust Soc Am 1989;86(3):1033–9. Harms MP, Melcher JR. Sound repetition rate in the human auditory pathway: representations in the waveshape and amplitude of fMRI activation. J Neurophysiol 2002;88(3):1433–50. Herdman AT, Lins O, Van Roon P, Stapells DR, Scherg M, Picton TW. Intracerebral sources of human auditory steady-state responses. Brain Topogr 2002;15(2):69–86. John MS, Dimitrijevic A, Picton TW. Efficient stimuli for evoking auditory steady-state responses. Ear Hear 2003;24(5):406–23. Joris PX, Schreiner CE, Rees A. Neural processing of amplitude-modulated sounds. Physiol Rev 2004;84(2):541–77. Kaukoranta E, Hamalainen M, Sarvas J, Hari R. Mixed and sensory nerve stimulations activate different cytoarchitectonic areas in the human primary somatosensory cortex SI. Neuromagnetic recordings and statistical considerations. Exp Brain Res 1986;63(1):60–6. Maurizi M, Almadori G, Paludetti G, Ottaviani F, Rosignoli M, Luciano R. 40-Hz steady-state responses in newborns and in children. Audiology 1990;29(6):322–8. Pantev C, Elbert T, Makeig S, Hampson S, Eulitz C, Hoke M. Relationship of transient and steady-state auditory evoked fields. Electroencephalogr Clin Neurophysiol 1993;88(5):389–96. Pantev C, Eulitz C, Elbert T, Hoke M. The auditory evoked sustained field: origin and frequency dependence. Electroencephalogr Clin Neurophysiol 1994;90(1):82–90. Pantev C, Roberts LE, Elbert T, Ross B, Wienbruch C. Tonotopic organization of the sources of human auditory steady-state responses. Hear Res 1996;101(1–2):62–74. Picton TW, Skinner CR, Champagne SC, Kellett AJ, Maiste AC. Potentials evoked by the sinusoidal modulation of the amplitude or frequency of a tone. J Acoust Soc Am 1987;82(1):165–78. Picton TW, John MS, Purcell DW, Plourde G. Human auditory steady-state responses: the effects of recording technique and state of arousal. Anesth Analg 2003;97(5):1396–402. Plourde G, Picton TW. Human auditory steady-state response during general anesthesia. Anesth Analg 1990;71(5):460–8. Rees A, Green GG, Kay RH. Steady-state evoked responses to sinusoidally amplitude-modulated sounds recorded in man. Hear Res 1986;23(2): 123–33. Reite M, Teale P, Zimmerman J, Davis K, Whalen J. Source location of a 50 msec latency auditory evoked field component. Electroencephalogr Clin Neurophysiol 1988;70(6):490–8. Reite M, Teale P, Goldstein L, Whalen J, Linnville S. Late auditory magnetic sources may differ in the left hemisphere of schizophrenic patients. Arch Gen Psychiatry 1989;46:565–72.
117
Reite M, Sheeder J, Teale P, Richardson D, Adams M, Simon J. Meg based brain laterality: sex differences in normal adults. Neuropsychologia 1995;33(12):1607–16. Reite M, Sheeder J, Teale P, Adams M, Richardson D, Simon J, Jones RH, Rojas DC. Magnetic source imaging evidence of sex differences in cerebral lateralization in schizophrenia. Arch Gen Psychiatry 1997;54: 433–40. Rojas DC, Teale P, Sheeder J, Simon J, Reite M. Sex-specific expression of heschl’s gyrus functional and structural abnormalities in paranoid schizophrenia. Am J Psychiatry 1997;154(12):1655–62. Ross B, Borgmann C, Draganova R, Roberts LE, Pantev C. A highprecision magnetoencephalographic study of human auditory steadystate responses to amplitude-modulated tones. J Acoust Soc Am 2000; 108(2):679–91. Ross B, Picton TW, Pantev C. Temporal integration in the human auditory cortex as represented by the development of the steady-state magnetic field. Hear Res 2002;165(1–2):68–84. Sarvas J. Basic mathematical and electromagnetic concepts of the biomagnetic inverse problem. Phys Med Biol 1987;1:11–22. Schoonhoven R, Boden CJR, Verbunt JPA, De Munck JC. Sources of meg and eeg auditory steady-state responses. In: Nowak H, Haueisen J, Geissler F, Huonker R, editors. Biomag 2002: proceedings of the 13th international conference on biomagnetism. Berlin: VDE Verlag; 2002. p. 116–8. Shallop JK, Osterhammel PA. A comparative study of measurements of SN-10 and the 40/sec middle latency responses in newborns. Scand Audiol 1983;12(2):91–5. Stapells DR, Linden D, Suffield JB, Hamel G, Picton TW. Human auditory steady state potentials. Ear Hear 1984;5(2):105–13. Stapells DR, Galambos R, Costello JA, Makeig S. Inconsistency of auditory middle latency and steady-state responses in infants. Electroencephalogr Clin Neurophysiol 1988;71(4):289–95. Suzuki T, Kobayashi K. An evaluation of 40-Hz event-related potentials in young children. Audiology 1984;23(6):599–604. Teale P, Sheeder J, Rojas DC, Walker J, Reite M. Sequential source model of the m100 exhibits inter-hemispheric asymmetry. NeuroReport 1998; 9:2647–52. Teale P, Reite M, Rojas DC, Sheeder J, Arciniegas D. Fine structure of the auditory m100 in schizophrenia and schizoaffective disorder. Biol Psychiatry 2000;48(11):1109–12. Teale P, Carlson J, Rojas D, Reite M. Reduced laterality of the source locations for generators of the auditory steady-state field in schizophrenia. Biol Psychiatry 2003;54(11):1149–53. Torrence C, Compo GP. A practical guide to wavelet analysis. Bull Am Meteorol Soc 1998;79:61–78. Witelson SF, Pallie W. Left hemisphere specialization for language in the newborn: neuroanatomical evidence of asymmetry. Brain 1973;96: 641–6.