A A A Volume : 44 Part : 2 ERP components analysis of selective attention to auditory signals in meaningful or meaningless noise using adaptive correlation filterTakahiro Tamesue 1Yamaguchi University 2-16-1, Tokiwadai, Ube, Yamaguchi, JapanABSTRACT Open o ffi ces that make e ff ective use of limited space and encourage dialogue, interaction, and collaboration among employees, are becoming an increasingly. However, productive work-related conversation might actually decrease the performance of other employees within earshot more so than other random, meaningless noises. On the other hands, it is well known that the Event-Related Potential (ERP) in the brain wave elicited by internal or external stimuli are related to the operation of selective attention. The present experiment was designed to determine the e ff ects of meaningfulness of external noise on ERP in the auditory odd-ball paradigms. First, in order to decide on a template of adaptive correlation filter, multivariate analysis such as the Principal Component Analysis (PCA) for ERP components was performed. Next, performance of algorithm on adaptive correlation filter for estimating average waveform of ERPs was evaluated. Furthermore, di ff erences in the ERP components due to the meaningfulness of external noise were examined.1. INTRODUCTIONOpen o ffi ces that make e ff ective use of limited space and encourage dialogue, interaction, and collaboration among employees, are becoming an increasingly. However, productive work-related conversation might actually decrease the performance of other employees within earshot more so than other random, meaningless noises. When carrying out intellectual activities involving memory or arithmetic tasks, it is common for external noise to increase levels of subjective annoyance, which can lead to a decline in performance. This tendency is stronger in response to meaningful noise, such as music and conversation, than for meaningless noise, such as the sound of tra ffi c, and heating, ventilating and air-conditioning noise. Hence, in designing a comfortable sound environment, it is important to understand the relationship between not only the measurable aspects of external noise, such as the sound pressure level, but also the qualitative aspects, such as the degree of meaningfulness of the external noise, and the subjective experience of annoyance. On the other hand, it is well known that the transient event related potentials (ERPs) elicited by internal or external stimuli in the brain wave are related to the operation of selective attention [1]. Previous studies have discussed how to evaluate event related potentials during intellectual task under meaningful or meaningless noise [2] [3]. In this study, the adaptive correlation filter (ACF) [4] is introduced in addition to the1 tamesue@yamaguchi-u.ac.jpa slaty. inter.noise 21-24 AUGUST SCOTTISH EVENT CAMPUS O ¥, ? GLASGOW conventional additional averaged method as a method to extract event related potentials more clearly. The present experiment was designed to determine the e ff ects of meaningfulness of external noise on ERP in the auditory odd-ball paradigms. First, in order to decide on a template of adaptive correlation filter, multivariate analysis such as the Principal Component Analysis (PCA) for ERP components was performed. Next, performance of algorithm on adaptive correlation filter for estimating average waveform of ERPs was evaluated. Furthermore, di ff erences in the ERP components due to the meaningfulness of external noise were examined.2. OUTLINE OF EXPERIMENTPsychophysiological experiments were conducted to determine the e ff ects of the meaningfulness of external noise on selective attention to auditory stimuli by examining di ff erences in brain event related potentials during the completion of the repetitive odd-ball paradigm. The outline of the experiments was as follows.2.1. Participants A total of 8 students with normal hearing participated in the experiment.2.2. Odd-ball paradigm The odd-ball paradigm is typically used to examine selective attention and information processing capacity [1]. In this task, subjects detect and respond to rare target events embedded in a series of repetitive events. Thus, to complete the odd-ball task it is necessary to regulate attention to a stimulus. In the auditory odd-ball paradigm, the common non-target stimulus (“frequent”) was a 1,000 [Hz] tone burst. The target stimulus was 2,000 [Hz] tone burst (“rare”) with an occurrence probability of 20 [%]. Both stimuli were presented binaurally at 60 [dB], and 120 [ms] duration (includng 10 [ms] rise-fall time and 100 [ms] plateau). The frequent-rare sequence was randomly presented with an inter-stimulus interval of 2 [s]. The subjects task was only to count the “rare” stimuli for approximately 10 [min].2.3. External noise The following external noises with di ff erent degree of meaningfulness, were employed as examples of typical indoor noises.(a) Meaningless noisePseudo voice-noise from a CD that was originally produced for the evaluation and fitting of hearing aids (TY-89) [5] was used as meaningless noise.(b) Meaningful noise 1Multi-talker noise from a CD for the evaluation and fitting of hearing aids (TY-89) [5], was used as meaningful noise 1.(c) Meaningful noise 2Male speech, produced by deleting handclaps, sound e ff ects, and music, etc. from commercially available speech tapes, was used as meaningful noise 2.For practical reasons, the energy-mean value of the sound pressure level of the above external noises was adjusted to approximately 50 [dB]. In addition, the following conditions were tested. -15-15-10-10-5-5Amplitude [ µ V]Amplitude [ µ V]005510101515(i) ACF (ii) Averaging(i) ACF (ii) Averaging20 0 100 200 300 400 50020 0 100 200 300 400 500Time [ms]Time [ms](a) Meaningless noise(b) Meaningful noise 1-15-15-10-10-5-5Amplitude [ µ V]Amplitude [ µ V]005510101515(i) ACF (ii) Averaging(i) ACF (ii) Averaging20 0 100 200 300 400 50020 0 100 200 300 400 500Time [ms]Time [ms](c) Meaningful noise 2(d) No external noiseFigure 1: Event related potentials for target stimulus ( L N : 50 [dB], Electrode: Fz)(d) No external noise2.4. Measurements Participants were seated in a sound-attenuated electrically shielded room. The auditory signal was generated by a CD player and presented through loud speaker. The electroencephalogram (EEG) was recorded from 20 locations ( Fp1, Fpz, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6 O1, O2 ) of scalp based on 10-20 system with Ag / Ag Cl electrodes of which impedance was held below 10 [k Ω ]. Electrodes were referenced to linked earlobes, and the ground electrode was placed on the midforehead electrode (Fpz). The electrooculogram (EOG) was recorded from an electrode located at the supra-orbital ridge of the right eye and referenced to the linked earlobes. The electroencephalogram and electrooculogram signals were amplified with a bandpass filter of 0.01 to 30 [Hz], and recorded with 16 [bit] quantization level at sampling rate of 1 [kHz], continuously. The event related potentials for the responses to the “rare” and “frequent” stimuli were synchronously averaged to enhance the evoked signal and suppress the background brain activity. -15-10-5Amplitude [ µ V]05101520TemplateTarget Non-target250 100 200 250 300 400 500Time [ms]Figure 2: Template for adaptive correlation filter, as a di ff erence in event related potentials between “rare” and “frequent” stimuli ( L N : 50 [dB], Electrode: Fz)3. EFFECT ON EVENT RELATED POTENTIALS OF MEANINGFULNESS OF THE EXTERNAL NOISE3.1. Conventional additional averaged method It is well established that event related potentials elicited by internal or external stimuli, can be measured using the electroencephalogram. A waveform of event related potentials after stimulus- triggered averaging to “frequent” and “rare” stimuli, was individually calculated on each electrode position under each external noise condition. As an example of results after stimulus-triggered averaging to “rare” stimuli, conventional averaged wave forms of event related potentials on the electrode position Fz, under the meaningless noise, meaningful noise 1, meaningful noise 2 and no external noise, with sound pressure level 50 [dB], are shown by blue broken line in Figure 1 (a), (b), (c) and (d). In the case of no external noise (Figure 1 (d)), there is a a positive peak occurring around 300 [ms] after presentation of stimuli. A socalled P300 component is thought to reflect the resolution of uncertainty or the perceptual decision that an expected signal has occurred. These components are related to selective attention and working memory. On the contrary, the P300 component is unclear under the meaningful noise 2 (Figure 1 (c)).3.2. Adaptive correlation filter Techniques such as principal component analysis (PCA) [6] use correlational structure of an event related potentials data set to define a set of components, may sometimes be useful for identifying latent event related potentials components. Similar to the previous study [3], principal components associated with P300 were extracted. Furthermore, an Adaptive Correlation Filter (ACF) [4] is introduced in addition to the conventional additive average method as a method to extract event related potentials more clearly. Based on detection of latency by correlation, samples of the electroencephalogram are cross correlated against template with a length of T and shifted amounts of time ∆ t at correlation coe ffi cient is maximum. For practical reasons, di ff erence in event related potentials between “rare” and “frequent” stimuli in the case of no external noise, was emploed as a template. Using the results of principal component analysis for event related potentials, length of the template, T , was adjusted to 250 [ms] from 250 to 500 [ms], ∆ t was adjusted to 125 [ms]. As an example of averaged wave forms of event related potentials with adaptive correlation filter, after stimulus-triggered averaging to “rare” stimuli, on the electrode position Fz, under the meaningless noise, meaningful noise 1, meaningful noise 2 and no external noise, with sound pressure level 50 [dB], are shown by red solid line in Figure 1 (a), (b), (c) and (d). We found reliable di ff erences in 0.30.30.250.25Relative FrequencyRelative Frequency0.20.20.150.150.10.10.050.0500-125 -100 -75 -50 -25 0 25 50 75 100 125-125 -100 -75 -50 -25 0 25 50 75 100 125Time [ms] (a) Meaningless noiseTime [ms] (b) Meaningful noise 10.30.30.250.25Relative FrequencyRelative Frequency0.20.20.150.150.10.10.050.0500-125 -100 -75 -50 -25 0 25 50 75 100 125-125 -100 -75 -50 -25 0 25 50 75 100 125Time [ms] (c) Meaningful noise 2Time [ms] (d) No external noiseFigure 3: Relative frequency distibution of di ff erence in latency ( L N : 50 [dB], Electrode: Fz)the P300 event related potentials between the external noise condition. These results indicate that attention to the “rare” stimulus was influenced by the degree of meaningfulness of the external noise during completion of auditory cognitive tasks. As an example of relative frequency distibution of latency di ff erence in adaptive correlation filter, on the electrode position Fz, under the meaningless noise, meaningful noise 1, meaningful noise 2 and no external noise, with sound pressure level 50 [dB], are shown in Figure 2 (a), (b), (c) and (d). It can be seen that the variation of the latency is larger under meaningless / meaningful noise than in the case no external noise. As an example of relative frequency distibution of correlation coe ffi cient in adaptive correlation filter, on the electrode position Fz, under the meaningless noise, meaningful noise 1, meaningful noise 2 and no external noise, with sound pressure level 50 [dB], are shown in Figure 3 (a), (b), (c) and (d). As the degree of meaningfulness of the external noise increases, the proportion of occurrences of high correlation coe ffi cients decreases.4. CONCLUSIONThis study focused on the e ff ects of the meaningfulness of external noise. We examined the e ff ects of meaningful noise and meaningless noise on physiological activity while carrying out auditory cognitive tasks. Specifically, the P300 components of the event related potentials elicited by the auditory odd-ball paradigms, were measured using the electroencephalogram. In order to examined 110.80.8Relative FrequencyRelative Frequency0.60.60.40.40.20.2000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Correlation (a) Meaningless noiseCorrelation (b) Meaningful noise 1110.80.8Relative FrequencyRelative Frequency0.60.60.40.40.20.2000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Correlation (c) Meaningful noise 2Correlation (d) No external noiseFigure 4: Relative frequency distibution of correlation coe ffi cient ( L N : 50 [dB], Electrode: Fz)di ff erences in the event related potential components due to the meaningfulness of external noise, performance of algorithm for estimating average waveform of event related potentials such as adaptive correlation filter is evaluated. The adopted correlation filter method revealed significant di ff erences of the P300 component between meaningful and no external noise during the completion of repetitive auditory oddball paradigm. The results revealed that the degree of meaningfulness of the external noise had a strong influence on selective attention to auditory stimuli in cognitive tasks. In conclusion, in designing comfortable sound environments in spaces used for cognitive tasks, it is appropriate to consider not only the sound pressure level, but also meaningfulness of the external noise that is likely to be present.ACKNOWLEDGMENTThis study was partially supported by the Japan Society for the Promotion of Science, Grant-in-Aid for Scientific Research (C), No. 18K11502, 21K12086.REFERENCES[1] S.Sutton, M.Braren, J.Zubin, and E.R.John, Evoked-potential correlates of stimulus uncertainty. Science , 150 (3700), 1187–1188, (1965). [2] T.Tamesue, E ff ects of acoustical noise on selective attention to auditory and visual stimuli, performance and annoyance during intellectual task. Proceedings 42th International Congress and Exposition on Noise Control Engineering , 2013.[3] T.Tamesue, Investigation of Selective Attention to Auditory Cognitive Task under Meaningful or Meaningless Noise. Proceedings 42th International Congress and Exposition on Noise Control Engineering , 2017.[4] Woody CD, Characterrization of an adaptive filter for the analysis of variable latency neuroelectric signals. Med Biol Eng 5 , 539-554, 1967.[5] K. Yonemoto, Characteristics of cd for the evaluation of fitting condition with hearing aids (ty-89). Journal of Otolaryngology, Head and Neck Surgery , 11 (9), 1395–1401, 1995.[6] E. Donchin, E. F. He ffl ey, Multivariate analysis of ERP data: A tutorial review. In D. A. Otto(Ed.), Multidisciplinary perspectives in event-related brain potential research. Washington, DC: US Government Printing O ffi ce. 555-572, 1978. Previous Paper 523 of 808 Next