Separation of Auditory Evoked Responses to the Right- and Left-Ear

Inputs

S. Kuriki

, H. Kurumaya

, K. Tanaka

and Y. Uchikawa

Research Institute for Science and Technology, Tokyo Denki University, Inzai, Japan

Department of Science and Engineering, Tokyo Denki University, Hatoyama, Japan

1 INTRODUCTION

In the clinical examination of auditory cortical

function, it is desirable to observe the evoked

responses separately to the right- and left-ear inputs

during binaural stimulation. Due to the dominance

of contralateral auditory pathway, it is expected that

the right ear response represents the function of the

left central auditory system and vice versa.

However, the response to the input of the same-ear

side is also included due to the conduction through

ipsilateral pathway. Difficulty exists how to

discriminate the mixed responses in the auditory

cortex into their input channels. From a view point

of basic science, separate observation of the two

ears’ responses to different acoustic stimuli such as

tones and speech sounds may be interesting to study

the hemispheric difference in the auditory function

(Zatorre and Belin, 2001). Here, we developed a

novel method for the separation of evoked responses

based on synchronous and asynchronous averaging

of signals.

2 PRINCIPLE

In the conventional averaging of evoked responses,

recorded signals, which are mixed with biological,

environmental and sensor noises, are averaged with

a trigger time-locked (synchronous) to the stimulus

epochs. This is based on the fact that noise signals

are asynchronous to the stimuli and thus attenuated

by averaging. Transient evoked responses are

usually phasic, being composed of several peaks

changing in polarity, e.g., P1, N1, P2, N2

components, of auditory response. It is expected that

such phasic responses are attenuated in amplitude

after averaging over many epochs if we use such

triggers that are asynchronous to the stimulus epochs.

This attenuation should be maximal when the jitter

of the trigger time extends to the whole response

period that may contain multiple peaks. For the

relevant response to be observed, synchronous

triggers are used besides the ongoing asynchronous

averaging, reducing the noise but not the signal.

We prepared two sets of time series of onset-

triggers in the simulation, where each set had

random onset-to-onset intervals varying in a period

of 500 ms. Figure 1a shows the distribution of the

onset intervals of the first trigger series, where bars

indicate the number of triggers included within time

bins of 50 ms width. The trigger onsets were

distributed uniformly in the 500 ms period, keeping

random intervals. It was confirmed that the second

trigger series prepared also had similar distribution.

(a)

(b)

Figure 1: (a) Distribution of the onset-to-onset intervals of

the triggers used in the simulation and (b) waveforms of

the signals averaged using synchronous and asynchronous

triggers.

As the signal, we used a waveform of auditory

evoked response consisting of several peak

components superposed with white noises. Signal

epochs of 500 ms length were generated repeatedly

at the onset of triggers of the first series. Averaging

Kuriki S., Kurumaya H., Tanaka K. and Uchikawa Y..

Separation of Auditory Evoked Responses to the Right- and Left-Ear Inputs.

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

was performed across 200 signal epochs using the

same trigger set and also using the second set of

triggers that were not time-locked to the signal

epochs. The result (Fig. 1b) showed clear waveform

of the original response for the synchronous

averaging, while no visible response was left for the

asynchronous averaging.

3 EXPERIMENTAL

We have carried out magnetoencephalograpy (MEG)

measurements of dichotic stimulation using a speech

sound /a/ to the right ear a 600-Hz pure tone to the

left ear of 500 ms length. The sound intensity was 70

dB SPL in both ears. Inter-stimulus intervals were

randomized between 250 and 750 ms, where the

right and left sounds were synchronized to two sets

of triggers having random onset times with no

correlation between them. Three male and a female

subjects of 22-25 years old participated in the

measurements. A whole head 122-channel magneto-

meter was used to measure MEG signals.

Averaging was performed across 200 epochs of

the recorded signals using the trigger set

synchronous to the tone epochs and also using the

other trigger set synchronous to the vowel epochs.

The averaged signals were filtered to 2-40 Hz offline.

The responses obtained at 6 channel sensors over the

right and left temporal areas were averaged to

represent the hemisphere response. The waveform of

the response was expressed as the root-mean-

squared magnitude.

4 RESULS AND DISCUSSION

Figure 2 shows the waveforms of the responses

obtained by synchronous averaging. The contra-

lateral vowel response was larger than ipsilateral

tone response in the left hemisphere, as expected.

However, the contralateral tone response was

reduced to a comparable magnitude to the ipsilateral

vowel response in the right hemisphere. The results

suggest that at short-interval stimulation (0.25-0.75

s) the vowel response in the left hemisphere does not

attenuate much, as compared with the tone response

and the responses in the right hemisphere.

MEG sensors in the temporal area can detect

exclusively the auditory cortical response in the right

or left hemisphere due to highly spatially localized

magnetic signals. So far, steady state responses that

are tagged with modulation frequency (Tononi et al.,

1998) have been used to discriminate the responses

to two-channel inputs. In addition to this, the present

method allows us to examine transient responses,

representing higher auditory functions, separately to

the speech and tone sounds in the two hemispheres.

It may be used in dichotic listening test in diagnosis

of auditory processing disorder (Moncrieff, 2006).

Figure 2: Waveforms of separated grand-mean responses

(n = 4) in the right and left hemispheres evoked by vowel

and pure tone stimuli, which were delivered to the right

and left ears, respectively.

REFERENCES

Moncrieff D. W., 2006. Identification of binaural

integration deficits in children with the Competing

Words Subtest: standard score versus interaural

asymmetry. Int J Audiol. 45:546-454.

Tononi G., Srinivasan R., Russel D. P., Edelman G. M.,

1998. Investigating neural correlates of conscious

perception by frequency-tagged neuromagnetic

responses. Proc. Natl. Acad. Sci. USA. 95:3198-3203.

Zatorre R. J., Belin P., 2001. Spectral and temporal

processing in human auditory cortex. Cereb Cortex.

11:946-953.