Date of Award
Doctor of Philosophy (PhD)
Professor Larry E. Roberts
Auditory evoked potentials recorded in the electroencephalogram consist of differentiable components that can be roughly classified as either transient or steady state responses. The latter, recorded at stimulus rates above about 3 Hz, have advantages for study of the auditory system. A large number of such responses can be collected rapidly to enhance the reliability of the resulting average; because the cortical sources of the response localize to the region of Heschl's gyrus, a picture of neural activity occuring in this region (primary auditory cortex, AI) is gained. An interesting feature of the steady-state response is that it reaches its peak amplitude at stimulation rates near 40 Hz. The experiments of this thesis investigated the mechanism and dynamics of the auditory steady-state response (SSR). The experiments were guided by the linear summation model which attributes the 40 Hz amplitude peack to the summation of transient "middle-latency" resposnes which contain a 40 Hz component and localize to Heschl's gyrus, overlapping the cortical sources of the SSR. Alternatively, it has been suggested that resonany properties of the auditory system may be responsible for the peak in SSE amplitude at 40 Hz.
Experiment 1 investigated how auditory evoked potentials change as stimulus rate increases from the transient to the steady state range. Stimuli consisting of 6 second long trains of stimulation at 1.5 Hz, 4 Hz or 13 Hz were presented to subjects in either the auditory, visual or somatosensory modality. The transient "on" response, "off response", and intervening responses evoked by stimuli in the train were recorded at each stimulus rate. Using the auditory recordings, frequency-amplitude functions were synthesized from the responses recorded at each of the stimulation rates. Only the 13 Hz response produced a frequency-amplitude characterisitc with a prominent peak in response amplitude at 40 Hz. A new signal processing procedure, based on the Hotelling T2 statistic but applied in a procedure similar to a spectrogram, was used to show dynamics in the steady-state response which were not observable in the time-domain averages.
In Experiment 2 a direct test of the linear summation model for the auditory steady-state response was performed. According to this model each of the indiviual stimuli in a train of steady-state stimulation induces an identical response referred to as the 'source transient'. Using a steady-state stimulus whose stimulation rate varied continuously from 10-50 Hz, I derived a 'source transient' using a deconvolution process (Gutschalk, Mase et al., 1999). The source transient was then used to reconstruct thet 10-50 Hz response assuming that linear summation was responsible for the entire response. The results did not find evidence of a nonlinear contribution to the generation of the SSR at any stimulus rate, and were therefore consistent with a linear model even at stimulus rates near 40 Hz. The source transient was also used to reconstruct a traditional 40 Hz steady-state response and compared to a recorded version of this response. The phase and amplitude of the reconstructed response accurately predicted the recorded response on the basis of individual subjects. The morphology of the group average 'source transient' was comparable with the middle-latency portion of the transient response recorded to a 1.5 Hz stimulus although its amplitude was reduced.
In Experiment 3 I applied a steady-state stimulus procedure to evaluate the plasticity of transient and steady-state components of the auditory evoked potential. Non-musician subjects were trained to discriminate small increases in the pitch of a 1 second long segment of a 40 Hz amplitude modulated tone with a carrier near 2 kHz. Using appropriate methods of signal processing, I separated N1 and P2 components of the transient response though to reflect activity in primary auditory cortex (A1). Training enhanced the amplitude of the P2 component in both hemispheres and the N1c component in the right hemisphere. The amplitude of the SSR was not enhanced by training; however, the phase of the response was modified for a time window commencing near the rising edge of the P2 component. Control measurements confirmed that the P2 and SSR were separate brain events. The conclusion contains a discussion about how auditory evoked potentials may reflect neural mecahnisms underlying remodeling of the auditory cortex by experience.
Overall the results of these studies suggest that linear summation gives an adequate description of the SSR when source transients are recorded at frequencies > 4Hz. My results are also consistent with neural generators for this response originating in Heschl's gyrus. Temporal properties of the steady-state response appear to be modifiable by training in adult non-musicians, although amplitude is more resistant to change. Exploration of a wider range of procedures may uncover additional dynamics in the SSR not detected here. These dynamics reflect fundmental mechanisms by which the brain encodes its sensory input.
Bosnyak, Daniel J., "Mechanisms and Dynamics of the Human Auditory Steady-State Response" (2003). Open Access Dissertations and Theses. Paper 851.