Effect of Attention and Noise on Echoic Memory as Indexed by the N1-Adaptation.

Effect of Attention and Noise on Echoic Memory as Indexed by the N1-Adaptation.

Ekenedilichukwu Tonia Osakwe

08.09.2021

There are numerous studies that support the notion that echoic memory is indexed by the adaptation of the N1 peak in auditory event related potentials (ERPs). Although the number research on the effects of parameters like noise and attention on the amplitude of the N1 is immense, to date there are no studies on the effect of these parameters on the adaptation of the N1. Here, I investigated the effect of noise and attention on the adaptation of N1, P2 and N1-P2. Secondary analysis was conducted on data collected from 33 participants in three conditions: passive recording condition (participant listen passively to stimulus while staring at a fixation cross); attention/oddball conditions (participant were task with counting the deviating tones); and noise condition where the tones are presented in white noise. Within each condition, two Stimulus onset intervals (SOI): 1.7 s and 3.5 were used in separate stimulus blocks and the ratio R = M1.7s / M3.5s was used as a dimensionless measure of adaptation. My results found no significant effect of noise an attention on the amplitudes and adaption of the N1, P2 and N1-P2. I propose that the lack of effect on the adaption of the ERPS might be due to noise and attention having a scaling effect on all of the amplitudes equally so that adaption lifetime is not affected. As this is the first study of its kind, further research will be needed to gain a better understanding of how adaptation is affected by these two factors.

Attention, Noise, N1-adaptation, auditory sensory memory

Participants
This project carries out secondary analysis on data from an EEG experiment with 33 human participants. The data was received from supervisor, Patrick May. The participants were all adult undergraduate and post graduate students at Lancaster University, with no self-reported hearing loss or neurological disorder. The experiment was approved by the research ethics procedures of the Department of Psychology, Lancaster University, and the participants provided written consent before the experiment began.

Equipment and Procedure for EEG measurements
Three dry electrodes were attached at locations: Fpz, Fz, Cz. Reference and ground electrodes were attached to the right ear lobe. For this report, only the data acquired from the Fz location was used as this is the channel that recorded the best ERPs for all the participants. The participants were directed to passively listen to stimuli while staring at a fixation cross and moving and blinking as little as possible. The stimuli comprised of 500-Hz pure tones with a duration of 100ms, including 10-ms linear onset and offset ramps. The stimuli were presented in blocks of 100 isochronous stimuli. The stimuli were presented binaurally via Sennheiser headphones using laboratory laptop and MATLAB interfaced with the Enobio EEG device in a soundproof chamber. Data was collected in three conditions: baseline passive recording condition (participant listen passively to stimulus while staring at a fixation cross); attention/oddball conditions (participant were task with counting the deviating tones); and noise condition where the tones are presented in white noise. Withing each condition, two Stimulus onset interval (SOI): 1.7 s and 3.5 were used in separate stimulus blocks. The order of experiments were randomised across the participants.
Data Analysis
The data was passband filtered at 1-30 Hz and sectioned into epochs of single trial data. To remove artefacts (e.g., due to blinking) 15% of epochs with the largest absolute amplitudes were removed. Single trial epochs was then averaged to reveal the ERP. The average ERP in a 100ms time window immediately preceding stimulus onset was calculated and subtracted from the whole ERP (baseline correction). The N1 is not the only peak that shows adaptation in auditory ERPS. Although many of the research on adaption is focused on the N1 peak, different researchers have looked at other auditory ERP peaks in relation to adaptations such as the P2 and P3 peaks. In fact, Lanting et al. (2013) found that the P2 was more very strongly affected by adaption than the N1. In addition, the peak-to-peak difference between the N1 and the P2 has been previously used to estimate adaptation in several studies as it provides a more reliable measure of activity in auditory cortex because as it has the advantage of not being dependent on the baseline activity which can be noisy (Lanting et al., 2013; Lavoie et al., 2008; Muller-Gass et al., 2008). Because of this, both the N1 and the P2 peaks were identified - the N1 was identified as the peak negativity at around 100ms and P2 peak positivity at around 200ms. The peak-to-peak difference between the N1 and the P2 was calculated and the N1 and P2 amplitude as well as the difference between the N1 and P2 amplitude was used to estimate the lifetime of adaptation. Statistical data analysis was conducted using Analysis of Variance (ANOVA). Specifically, three one-way (condition) and three two-way (SOI x condition) repeated measures ANOVAs was conducted of the N1, P2 and the difference between the N1 and P2 amplitudes and amplitude ratios respectively.

Calculating the lifetime of adaptation (τ)
The recovery time constant for adaptation is usually calculated by fitting an exponentially saturating function to peak amplitudes plotted across SOIs (Lu et al., 1992). This curve is characterized by  as well as by two other fitting parameters: asymptotic magnitude and crossing point on SOI axis. The parameter  determines the steepness of the magnitude curve: the smaller its value, the quicker the curve approaches the asymptote (i.e., levels out) as SOI is increased. The SOIs where this levelling out has occurred represent stimulation where the silent period between two consecutive stimuli is large enough for adaptation to have died away. Therefore,  expresses the lifetime of adaptation: with low values, the curve levels out to its maximum value quicker; with high values, the amplitude rises slower as a function of SOI, meaning that adaptation is strongly present in a larger range of SOIs.
For fitting the exponential function reliably, a large number of SOIs should be employed, and the largest SOI should measure approximately 10s to ensure that adaptation has died away. Coupled with the requirements of data quality (large number of stimulus repetitions), this means long measurement times. In this experiment, this was bypassed by noting that the ratio between the magnitudes measured at two different SOIs is proportional to . Expressing the magnitudes of the brain responses measured at SOIs 1.7 s and 3.5 s by M1.7s, and M3.5s, respectively, the ratio R = M1.7s / M3.5s was used as a dimensionless measure of  and adaptation lifetime. The smaller R is, the shorter adaptation lifetime is. R was calculated separately for each participant for each of the experimental conditions and for each SOI. In addition, R was also calculated separately for the N1 and P2 peaks as well as the difference between these peaks. Note that the actual adaptation lifetime cannot be estimated by the use of this method.

Results
18 participants’ data did not show identifiable ERP responses and were thus discarded from analysis. The ERPs obtained from the final sample of 15 were plotted as shown in Figure 1 for each participant. The means and standard deviations were then calculated for the identified N1, P2 and the difference between the N1 and P2 for each SOI and condition as shown in Table 1. Seeing as there is such a large variability across the conditions, it is predictable that no statistical differences were found by the ANOVA.

Lancaster University

data/r.csv

Osakwe2021

Emily Dreyer
Paige Durnall

Open

None

English

Data

LA1 4YF

Patrick May

MSC

Neuroscience, Neuropsychology

33 to start, 18 were removed so final number is 15

ANOVA