Influence of Workload on Auditory Evoked Potentials

in a Single-stimulus Paradigm

R. N. Roy

1,2

, A. Breust

, S. Bonnet

, J. Porcherot

, S. Charbonnier

, C. Godin

and A. Campagne

CEA LETI, Univ. Grenoble Alpes, Minatec Campus, Grenoble, France

Gipsa-Lab, Univ. Grenoble Alpes and CNRS, Grenoble, France

LPNC, Univ. Grenoble Alpes and CNRS, Grenoble, France

Keywords: Workload, Mental State Monitoring, ERP, Audition.

Abstract: Mental workload can be assessed via neurophysiological markers. Temporal features such as event related

potentials (ERPs) are one of those which are very often described in the literature. However, most of the

studies that evaluate their sensitivity to workload use secondary tasks. Yet potentials elicited by ignored

stimuli could provide mental state monitoring systems with less intrusive probing methods. For instance,

auditory probing systems could be used in adaptive driving or e-learning applications. This study evaluates

how workload influences auditory evoked potentials (AEPs) elicited by a single-stimulus paradigm when

probes are to be ignored. Ten participants performed a Sternberg memory task on a touchpad with three

levels of difficulty plus a view-only condition. In addition, they performed two ecological tasks of their

choice, one deemed easy (e.g. reading novels), and the other difficult (e.g. programming). AEPs were

elicited thanks to pure tones presented during the memory task retention period, and during the whole extent

of the external tasks. Performance and AEPs were recorded and analyzed. Participants’ accuracy decreased

linearly with increasing workload, whereas the difference in amplitude between the P3 and its adjacent

components, N2 and SW, increased. This reveals the relevance of this triphasic sequence for mental

workload assessment.

1 INTRODUCTION

1.1 Workload and Mental State

Monitoring

The aim of mental state monitoring (MSM) and

neuro-ergonomics is to evaluate an operator’s state

in order to better supply her/him with help,

information or safety measures. The applications are

numerous, ranging from gaming to education,

including driving and security. The new systems that

allow this assessment are called passive Brain

Computer Interfaces (pBCI; Zander et al., 2011),

biocybernetic systems, or physiologically attentive

user interfaces when they adapt their functionality to

the user’s covert state (Chen and Vertegaal, 2004). It

is well known that mental state modulations are

reflected by a variety of physiological signals,

including neurophysiological signals. Hence, passive

BCI systems use neural markers, such as spectral or

temporal features to classify a given mental state.

Most of those neural markers are derived from

electro-encephalographical (EEG) signals (Blankertz

et al., 2010; van Erp et al., 2012).

One of the mental states that are of major interest

to evaluate is workload. Mental workload has been

extensively documented and can be defined either as

the load in memory (i.e. number of items), the

number of tasks to be performed in parallel, and

more generally as a measure of the amount of

cognitive and attentional resources engaged in a

task. It is therefore considered to be close to task

difficulty (Gevins and Smith, 2007), and to depend

on each individual’s capabilities and the effort put in

performing the task (Cain, 2007). Therefore, a

classical finding is that performance (e.g. response

time and accuracy) decreases when workload

increases (e.g. Sternberg, 1966; Natani et al., 1981;

Gomarus et al., 2006).

1.2 EEG Markers of Workload

Several EEG features are known to react to an

increase in workload. Spectral features such as

power spectral density have been thoroughly

described in the literature (e.g. Gomarus et al., 2006;

Berka et al., 2007; Roy et al., 2013). For instance,

104

Roy R., Breust A., Bonnet S., Porcherot J., Charbonnier S., Godin C. and Campagne A..

Inﬂuence of Workload on Auditory Evoked Potentials in a Single-stimulus Paradigm.

DOI: 10.5220/0005235701040111

In Proceedings of the 2nd International Conference on Physiological Computing Systems (PhyCS-2015), pages 104-111

ISBN: 978-989-758-085-7

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

numerous indices have been created to better exploit

the variation in the alpha and beta bands. One of

such is the ratio of the theta activity at frontal sites

with the alpha activity at parietal sites (Gevins and

Smith, 2003; Holm et al., 2009).

Another commonly used feature is a temporal

one: event related potentials (ERPs). Indeed, a

stimulus, either from the task at hand or an external

probe, can be used to elicit a neural response. Most

studies that report variations of ERP components

due to workload variations use visual stimuli.

However, auditory or tactile ERPs should be

considered for real-life applications. Indeed, systems

that would make use of probes from other sensory

modalities than vision would be less intrusive, less

distracting and therefore less risky for the operator

and its task. For instance, in a driving situation,

ERPs elicited by tactile stimuli from the chair could

be a good means to evaluate the drivers’ workload

(Sugimoto and Katayama, 2013). In the same

manner, auditory probes can be good indicators of

one’s mental workload.

Auditory evoked potentials (AEPs) are typically

studied using oddball paradigms. The elicited

components include a typical triphasic sequence

(Smith et al., 1990): N2, P3 and Slow Wave (SW).

When stimuli are ignored, the P3 component is

anterior (P3a) and reflects an involuntary capture of

attention, whereas when stimuli are attended to, it

has a posterior distribution (P3b; Squires et al.,

1975; Näätänen and Gaillard, 1983; Halgren et al.,

1998; Strobel et al., 2008; Allison and Polich, 2008).

It is well known that the P3 component has an

amplitude reduction when workload increases, for

both visual and auditory probes (Natani and Gomer,

1981; Kok, 2001; Schultheis and Jameson, 2004;

Holm et al., 2009; Miller et al., 2011; Brouwer et al.,

2012). Earlier components such as the N1, P2 or the

N2 have also been reported to be sensitive to

variations in workload (Ullsperger et al., 2001;

Allison and Polich, 2008; Miller et al., 2011). As for

later slow waves, to our knowledge no effect of

workload has been reported on the negative

component appearing just after the P3.

All those studies on the impact of workload on

AEPs have been conducted using classical oddball

paradigms in which participants had to detect

(and/or count) a target infrequent item amongst

distractors or novel sounds. However, for real-life

applications of mental state monitoring systems, a

less intrusive and distracting probing method should

be used. That is to say that the use of a secondary

task should be avoided in order to keep the operator

focused on its primary task. Hence, Allison and

Polich (2008) have introduced the single-stimulus

paradigm for assessing mental workload in an

immersive environment in a less distracting way. In

this paradigm, there are no non-target stimuli, they

are replaced by silence, and only target stimuli are

presented, at irregular intervals. As the authors point

out, this is a stimulation method that is operationally

easy to implement. In their study, participants had to

either count or ignore these auditory stimuli while

playing a video game. The authors indicated

amplitude modulations for the P2, N2 and P3

components. However, the modulations were not the

same depending on the contrast. For instance, it

decreased when participants were playing in a

difficult condition vs. a medium one. But, it

increased from an easy one to a medium one.

1.3 Current Study

As we saw earlier, most of the studies that assessed

the impact of workload on AEPs have used

secondary tasks. However, this is less realistic to

implement for real-life applications. It seems that the

single-stimulus paradigm is the best way to elicit

AEPs by interfering as little as possible in the

operator’s task. However, to our knowledge, only

one study has paved the way to evaluating the

usability of potentials evoked by this paradigm for

mental workload assessment (Allison and Polich,

2008), and it gave no definite conclusion as to a

robust amplitude modulation of ERP components.

Therefore, these results need to be supported and

extended. We intend to go further by examining how

AEPs elicited by a single-stimulus paradigm are

influenced by workload for a laboratory task, the

Sternberg memory task, in which participants have

to memorize a varying number of items, and also for

work-related ecological tasks, i.e. reading and

computer programming. In order to do so, we used

pure tones of different frequencies, and we extracted

the amplitude and latency of the triphasic AEP

sequence N2, P3, SW, as well as the LPP elicited by

these probes. A potentially interesting new feature

was also evaluated, i.e. the difference in amplitude

between these adjacent components.

2 MATERIALS AND METHODS

2.1 Experimental Protocol

Ten healthy right-handed participants (7 males; m =

28.89 years, s.d. = 7.02 years) volunteered for the

experiment.

InfluenceofWorkloadonAuditoryEvokedPotentialsinaSingle-stimulusParadigm

105

Workload was manipulated using a modified

Sternberg task (Sternberg, 1966) and ecological

external tasks. The first task was performed on a

Windows Surface touchpad and was implemented in

C++ using Qt/qwt and Visual Studio C++. As for the

ecological tasks, they were performed by

participants on their work computer or on their desk.

In the modified Sternberg task, the 20 French

consonants were used as visual stimuli (vowels were

excluded to avoid chunking strategies). In addition,

for both the Sternberg task and the external tasks,

the auditory probes were six pure tones ranging from

750 to 2000 Hz with a 250 Hz step with a 100 ms

duration, including a 10 ms rise and a 10 ms fall.

They were presented binaurally using a Logitech PC

headset.

Figure 1: EEG electrode and acquisition system setup.

During the Sternberg task, participants had to

memorize a list of sequential consonants visually

presented on the touchpad screen. Then, a keyboard

was presented (Figure 2). The participants had to

retrieve the consonants and the order in which they

were presented, and answer as accurately as possible

by touching the screen. Three levels of workload

were considered, i.e. 3, 5 and 7 consonants to

memorize (easy, medium and high workload

respectively), as well as a ‘view-only’ condition, or

idle state, in which they only focused their attention

on the fixation point and had no item to memorize or

retrieve. All trials were pseudo-randomly presented.

The six auditory probes were presented during the

retention period, with an inter-tone interval of 2005

+ i * 1000 ms, i ranging from 0 to 5. Participants

performed a training session of 3 easy and 3 difficult

trials for 4.5 minutes. Then, they performed 3 trials

of the 4 conditions: easy, medium, hard and view-

only, for 9 minutes.

Then, the participants had to choose two tasks

amongst several ecological tasks: an easy and a

difficult one (as defined by us). For the easy task, 4

participants watched YouTube movies without the

sound or read funny stories on 9gag.com, 4 read

novels or newspapers and 2 surfed on the internet.

For the difficult task, 5 participants who are

computer scientists or students in computer science

wrote code on their computer, 2 read scientific

publications in English (not their mother tongue),

and the last 3 played difficult Sudoku grids online.

Each external task lasted for 15 minutes. In total, the

whole experiment lasted for 44 minutes.

2.2 Measures and Analyses

The accuracy of participants’ answers to the

implemented task was recorded, as well as their

EEG activity for all tasks using the Robik box

acquisition system (Filipe et al., 2011). EEG activity

Figure 2: Trial structure for the Sternberg memory task.

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

106

was recorded from four passive Ag/AgCl electrodes:

Oz, Pz, Cz and T10 positioned according to the 10-

20 system. These electrodes were fastened on a

double strap headband and kept moist with

physiological serum. The reference was set at Fpz

and the ground electrode at the right earlobe (Figure

1). The data were sampled at 390 Hz, band-pass

filtered between 1 and 20 Hz, and re-referenced to a

common average reference.

EEG analyses were only performed on the signal

acquired from the Cz electrode, as it proved relevant

for workload estimation from both early and late

components (Allison and Polich, 2008). ERPs

elicited by the auditory probes during the different

tasks were extracted by subtracting a 200 ms pre-

stimulation baseline to the 600 ms post-stimulation

signal. Trials with a maximum over 100 µV were

rejected. Absolute peak amplitude and latency were

extracted for the 4 following components: the N2,

negative deflection between 150 and 250 ms, the P3,

positive deflection between 220 and 350 ms, the

SW, negative deflection between 350 and 450 ms,

and the LPP, positive deflection between 400 and

600 ms. The differences in amplitude between P3

and N2 (P3N2), P3 and SW (P3SW), and LPP and

SW (LPPSW) were also extracted.

Figure 3: Accuracy of participants’ answer in the

Sternberg task depending on workload condition –

Average and standard deviation across participants.

To statistically assess workload’s impact on both

accuracy and AEPs (amplitude and latency), we

performed repeated measure ANOVAs with Tukey

post-hoc tests. The analyses of accuracy were

performed using Statistica, and all the EEG analyses

were performed using Matlab. Performance in the

Sternberg task was only evaluated for the conditions

in which an answer was expected, i.e. the easy,

medium and hard conditions. As for EEG data

analysis, in the Sternberg task, we had 4 levels of

workload: easy, medium, hard and view-only. Only

trials for which a correct answer was given were

kept, in order to effectively evaluate the impact of an

increasing number of items in memory. For the

external tasks, we had 2 workload levels: easy and

hard.

3 RESULTS

3.1 Task Performance

Participants were less accurate when workload

increased (p<0.001), and this effect was linear

(linear polynomial p<.01; quadratic polynomial n.s.;

Figure 3). We can observe a large variability

between participants, mostly for the medium and

hard conditions.

3.2 Auditory Evoked Potentials

Grand averages across participants of the ERPs

elicited by the pure tones in the Sternberg and the

external tasks depending on workload condition are

respectively given in Figures 4 and 5. It should be

noted that there were two peaks to the P3 component

for the different conditions, but for the view-only

one.

Although the averaged signal shows modulations

of components’ amplitude depending on workload

condition for both the implemented task and the

external tasks (e.g. increased N2 and SW

amplitudes, decreased P3 amplitude), these

differences, as well as latency differences were not

significant. This is most certainly due to too much

variance between participants resulting in a levelling

at the group level. This variability is illustrated by

Figure 6 which displays the number of participants

that show a significant difference in voltage across

trials depending on time and condition comparison

(e.g. E vs. M: Easy vs. medium difficulty levels).

Indeed, on our total number of ten subjects, only a

maximum of three participants have a significant

difference congruent in time. Nevertheless, we can

see that the time periods of the N2, P3, SW and LPP

components are somewhat relevant at the participant

level.

Interestingly, it so appeared that the differences

in amplitude between adjacent components were

more robust to this inter-participant variance.

Indeed, for the Sternberg task, the P3N2

significantly increased with workload (F(3,27) =

InfluenceofWorkloadonAuditoryEvokedPotentialsinaSingle-stimulusParadigm

107

8.39 , p<.001), and this effect was linear between the

easy, medium and hard conditions (linear

polynomial p<.01; quadratic polynomial n.s.). The

same workload effect was observed on the P3SW

(F(3,27) = 4.52, p<.05). It was also linear between

the easy, medium and hard conditions (linear

polynomial p<.05; quadratic polynomial n.s.).

Figure 4: Impact of workload condition on the AEPs at the

Cz electrode during the Sternberg task. Data smoothed

using a 5-sample moving average.

Figure 7 displays the difference waveforms between

the hard and easy, and the medium and easy

conditions. It illustrates well the increase in

difference for these adjacent components with

increasing workload. Figure 8 details those

amplitude differences for the Sternberg task through

box plots. For the LPPSW, no significant difference

was observed.

Figure 5: Impact of workload condition on the AEPs at the

Cz electrode during the ecological tasks. Data smoothed

using a 5-sample moving average.

As regards the external tasks, although the signal at

the Cz electrode seemed greatly modulated by

workload, the impact of this factor was in fact

smaller on peak amplitudes, and was not significant.

Figure 6: Number of participants that show a significant

difference in voltage across trials depending on condition

comparison and time. E: easy; M: medium, H: hard, V:

view-only, Ext: external tasks.

Figure 7: Difference waveforms at the Cz electrode for the

Sternberg task. Data smoothed using a 5-sample moving

average.

4 DISCUSSION

The aim of this study was to assess the impact of

workload on auditory evoked potentials elicited by a

single-stimulus paradigm with probes ignored by the

participants. Workload was efficiently manipulated

using a Sternberg paradigm with decreasing

performance when the task increased in difficulty.

0 100 200 300 400 500 600

-6

-4

-2

Time (ms)

Amplitude (

Easy

Medium

Hard

View only

0 100 200 300 400 500 600

-2

-1

Time (ms)

Amplitude (

Easy

Hard

Time (ms)

Comparison

0 100 200 300 400 500

E vs M

E vs H

M vs H

V vs E

V vs M

V vs H

Ext.

0 100 200 300 400 500 600

-6

-5

-4

-3

-2

-1

Time (ms)

Amplitude (

Hard - Easy

Medium - Easy

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

108

Figure 8: Box plots of the amplitude differences between

the P3 and N2 components, and the P3 and SW

components, at the Cz electrode for the Sternberg task.

Average across participants.*p<0.05.

The main findings of this study are a significant

modulation of the N2, P3 and SW triphasic AEP

sequence for the Sternberg task. Indeed, the

difference in amplitude between the P3 and each of

the two adjacent components N2 and SW increased

with workload. This phenomenon has never been

described in the literature and could provide an

interesting feature for mental workload estimation as

it seems more robust to inter-participant variability

than components’ amplitude.

Indeed, no significant effect of workload was

found for these amplitudes, nor for the latencies of

the components at the group level. Although not

significant, our components’ amplitude results are in

line with the work of Allison and Polich (2008).

They introduced the use of the single-stimulus

paradigm and workload impact on ignored probes,

and reported larger components for the view-only

condition. They also indicated that an increase in

components’ amplitude could be observed when

workload increases, as we found for the N2 and SW

components. This also explains the increase in

amplitude difference for the P3N2 and the P3SW.

Also, similarly, they did not report significant

modulations of AEPs’ amplitude between the easy

and medium, and easy and hard conditions with their

experimental paradigm. Lastly, the absence of any

significant modulation of our triphasic sequence for

the ecological tasks may be due to a bad choice of

ecological tasks, or to an insufficient engagement

from the participants. These issues are critical for

real-life experimentations.

However, this study is just a preliminary study

and the results need to be further examined and

confirmed using more participants, as well as more

trials per condition and additional ecological tasks.

Also, a problem of variable lag between trigger and

sound release due to the use of the touchpad has

appeared. It could have brought more variance and

therefore reduced the effects. Next time, we will

record the audio signal along with the EEG signal to

realign our ERPs. That being said, these significant

differences in amplitude might appear when

performing classification at the subject level.

Therefore, the next step should be to try and

estimate each participant’s mental workload using

these amplitudes, as well as the P3N2 and P3SW

differences and compare their efficiency. This

should be done for laboratory-type tasks as well as

for ecological tasks, as we have done in this study.

Our study brings new light on the use of AEPs,

as well as the single-stimulus paradigm for mental

state monitoring. Robust features such as differences

in amplitude could be used for workload assessment

in a non-intrusive way by probing operators with

pure tones irrelevant to their task at hand.

5 CONCLUSIONS

This study fits into the mental state monitoring

growing research environment. Mental workload

assessment is a new challenge that can be tackled by

evaluating the relevance of several neuro-

physiological features, such as auditory evoked

potentials. Our results show that the amplitude of

these potentials elicited by pure tones in a single-

stimulus fashion are modulated by workload for

laboratory-type tasks as well as ecological tasks,

although not significantly at the group level.

However, the differences in amplitude between the

adjacent components of the triphasic AEP sequence

N2, P3 and SW were significantly modulated for our

laboratory task. These promising results should be

taken to the next step by comparing their relevance

with other features using classification algorithms.

The use of more electrodes as well as other

recording modalities should also be considered to

improve mental workload assessment. Finally, with

InfluenceofWorkloadonAuditoryEvokedPotentialsinaSingle-stimulusParadigm

109

the aim to get closer to real-life implementation, a

thorough ecological task battery setup should be

designed to better ascertain workload modulation in

work settings.

REFERENCES

Allison, B. Z., and Polich, J. (2008). Workload assessment

of computer gaming using a single stimulus event-

related potential paradigm. Biological Psychology,

77(3), pp. 277-283.

Berka, C., Levendowski, D. J., Lumicao, M. N., Yau, A.,

Davis, G., Zivkovic, V. T., Olmstead, R. E., et al.

(2007). EEG correlates of task engagement and mental

workload in vigilance, learning, and memory tasks.

Aviation, Space, and Environmental Medicine, 78,

231-244.

Blankertz, B., Tangermann, M., Vidaurre, C., Fazli, S.,

Sannelli, C., Haufe, S., Maeder, C., Ramsey, L.,

Sturm, I., Curio, G., and Müller, K.-R.. (2010). The

Berlin Brain–Computer Interface: Non-Medical Uses

of BCI Technology. Frontiers in Neuroscience, vol. 4.

Brouwer, A.-M., Hogervorst, M. A. , van Erp, J. B. F.,

Heffelaar, T., Zimmerman, P. H. and Oostenveld, R.

(2012). Estimating workload using EEG spectral

power and ERPs in the n-back task. Journal of Neural.

Enineering, 9(4).

Cain, B. (2007). A review of the mental workload

literature. Defense Research and Development,

Toronto, Canada, [Online]. Available:

http://ftp.rta.nato.int/public//PubFullText/RTO/TR/RT

O-TR-HFM-121-PART-II///TR-HFM-121-Part-II-

04.pdf.

Chen, D. and Vertegaal, R.(2004). Using mental load for

managing interruptions in physiologically attentive

user interfaces. In Human factors in computing

systems, pp. 1513-1516.

van Erp,, J., Lotte, F., and Tangermann, M. (2012). Brain-

Computer-Interfaces: Beyond Medical Applications.

Computer, vol. 45, pp.26-34.

Filipe, S., Charvet, G., Foerster, M., Porcherot, J., Bêche,

J.-F., Bonnet, S., Audebert, P., Régis, G., Zongo, B.,

Robinet, S., Condemine, C., Mestais, C., Guillemaud,

R. (2011). A wireless multichannel EEG recording

platform. In Conf. Proc. of the IEEE Eng. in Med. and

Biol. Soc., Boston, Massachusetts USA, pp. 6319-

6322.

Gevins, A., and Smith, M. E. (2003). Neurophysiological

measures of cognitive workload during human-

computer interaction. Theoretical Issues in

Ergonomics Science, 1, 113-131.

Gevins, A., and Smith, M. E. (2007).

Electroencephalography (EEG) in Neuroergnomics. In

Neuroergonomics: The brain at work, R. Parasuraman,

R and M. Rizzo Eds. New York, NY: Oxford.

Gomarus, H. K., Althaus, M., Wijers, A. A., and

Minderaa, R. B. (2006). The effects of memory load

and stimulus relevance on the EEG during a visual

selective memory search task: An ERP and ERD/ERS

study. Clinical Neurophysiology, 117, 871-884.

Halgren, E., Marinkovic, K., and Chauvel, P. (1998).

Generators of the late cognitive potentials in auditory

and visual oddball tasks. Electroencephalography and

clinical Neurophysiology. 106, pp. 156-164.

Holm, A., Lukander, K., Korpela, J., Sallinen, M. and

Müller, K. M. I. (2009). Estimating Brain Load from

the EEG. TheScientificWorldJOURNAL, vol. 9, pp.

639-651.

Kok, A. (2001). On the utility of P3 amplitude as a

measure of processing capacity. Psychophysiology, 38,

557–577.

Miller, M. W., Rietschel, J. C., McDonald, C. G., and

Hatfield, B. D. (2011). A novel approach to the

physiological measurement of mental workload.

International Journal of Psychophysiology, 80, 75-78.

Näätänen, R. Gaillard, A. W. K., and Varey, C. A. (1981).

Attention effects on auditory EPs as a function of

inter-stimulus interval. Biological Psychology, 13, pp.

173-187.

Natani, K., and Gomer, F. E. (1981). Electrocortical

activity and operator workload: A comparison of

changes in the electroencephalogram and in event-

related potentials. (McDonnell Douglas Technical

Report E2427). Long Beach, CA: McDonnell Douglas

Corporation.

Roy, R. N., Bonnet, S., Charbonnier, S. and Campagne, A.

(2013). Mental fatigue and working memory load

estimation: Interaction and implications for EEG-

based passive BCI. In Conf. Proc. of the IEEE Eng.

Med. Biol. Soc., Osaka, Japan, pp. 6607-6610.

Schultheis, H. and Jameson, A. (2004) Assessing

Cognitive Load in Adaptive Hypermedia Systems:

Physiological and Behavioral Methods. Lecture Notes

in Computer Science, 313, 225-234.

Smith, M. E., Halgren, E., Sokolik, M., Baudena, P.,

Musolino, A., Liegeois-Chauvel, C., and Chauvel, P.

(1990). The intracranial topography of the P3 event-

related potential elicited during auditory oddball.

Electroencephalography and clinical neuro-

physiology, 76, pp. 235-248.

Squires, N. K., Squires, K. C., and Hillyard, S. A. (1975).

Two varieties of long-latency positive waves evoked

by unpredictable auditory stimuli in man.

Electroencephalography and clinical neuro-

physiology, 38, pp. 987-401.

Sternberg, S. (1966). High-speed scanning in human

memory. Science, vol. 153, pp. 652-654.

Strobel, A., Debener, S., Sorger, B., Peters, J. C.,

Kranczioch, C., Hoechstetter, K., Engel, A. K.,

Brocke, B., and Goebel, R. (2008). Novelty and target

processing during an auditory novelty oddball: A

simultaneous event-related potential and functional

magnetic resonance imaging study. NeuroImage, 40,

869-883.

Sugimoto, F. and Katayama, J. (2013). Somatosensory P2

reflects resource allocation in a game task: assessment

with an irrelevant probe technique using electrical

probe stimuli to shoulders. International Journal of

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

110

Psychophysiology, 87, 200-204.

Ullsperger, V., Freude, G. and Erdmann, U. (2001).

Auditory probe sensitivity to mental workload changes

- an event-related potential study. International

Journal of Psychophysiology, 40, pp. 201-209.

Zander, T. O. and Kothe, C. (2011). Towards passive

braincomputer interfaces: applying brain-computer

interface technology to human-machine systems in

general. Journal of Neural. Enineering., 8(2):025005.

InfluenceofWorkloadonAuditoryEvokedPotentialsinaSingle-stimulusParadigm

111