Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Prestimulus vigilance predicts response speed in an easy visual discrimination task

Prestimulus vigilance predicts response speed in an easy visual discrimination task Background: Healthy adults show considerable within-subject variation of reaction time (RT) when performing cognitive tests. So far, the neurophysiological correlates of these inconsistencies have not yet been investigated sufficiently. In particular, studies rarely have focused on alterations of prestimulus EEG-vigilance as a factor which possibly influences the outcome of cognitive tests. We hypothesised that a low EEG-vigilance state immediately before a reaction task would entail a longer RT. Shorter RTs were expected for a high EEG-vigilance state. Methods: 24 female students performed an easy visual discrimination task while an electroencephalogram (EEG) was recorded. The vigilance stages of 1-sec-EEG-segments before stimulus presentation were classified automatically using the computer-based Vigilance Algorithm Leipzig (VIGALL). The mean RTs of each EEG-vigilance stage were calculated for each subject. A paired t-test for the EEG-vigilance main stage analysis (A vs. B) and a variance analysis for repeated measures for the EEG-vigilance sub-stage analysis (A1, A2, A3, B1, B2/3) were calculated. Results: Individual mean RT was significantly shorter for events following the high EEG-vigilance stage A compared to the lower EEG-vigilance stage B. The main effect of the sub-stage analysis was marginal significant. A trend of gradually increasing RT was observable within the EEG-vigilance stage A. Conclusion: We conclude that an automatically classified low EEG-vigilance level is associated with an increased RT. Thus, intra-individual variances in cognitive test might be explainable in parts by the individual state of EEG- vigilance. Therefore, the accuracy of neuro-cognitive investigations might be improvable by simultaneously controlling for vigilance shifts using the EEG and VIGALL. Introduction High intra-individual variance of behavioural measures Typically, studies in cognitive neuroscience implement leads to biased evaluations of cognitive processing, both paradigms, e.g. cognitive performance tasks, in which within healthy subjects as well as patient cohorts. Pro- participants respond to randomly presented sensory sti- longation of an experimental paradigm is a frequently muli. By comparing averages of stimulus-locked used and common method to reduce intra-individual responses, such as reaction time (RT) or error rate (ER), variance for obtaining a more reliable measure of RT valuable information on cognitive processing can be [7]. Nevertheless, the amount of experimental trials can gained. Beyond the variability between different subjects be limited, e.g. due to reduced physical and mental con- dition of patients or older subjects. Thus, intra-indivi- (inter-individual variability), responses of the same sub- ject vary crucially (intra-individual variability) across dual variability should be minimised by controlling for experiments [1]. Previous studies report that inter-indi- potential covariates that directly influence RT and ER. vidual differences in RT are associated with gender, age Therefore, we focused on the fluctuating state of [2] and neurological alterations [3-6]. wakefulness, as we hypothesized that the level of alert- ness crucially impacts individual performance during an * Correspondence: Hubertus.Himmerich@medizin.uni-leipzig.de experimental procedure. For examining our hypothesis, Department of Psychiatry and Psychotherapy, University of Leipzig, we adhere to the EEG-vigilance concept, which unfortu- Semmelweisstr. 10, 04103 Leipzig, Germany nately overlaps with other concepts, e.g. alertness, Full list of author information is available at the end of the article © 2011 Minkwitz et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 2 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 attention and arousal [8]. We use the term vigilance to Nevertheless, examination of EEG-vigilance based refer to different levels of brain function on the sleep- variability in RT tasks has not been used on a single- wake spectrum as they are empirically assessable by trial basis so far due to methodological difficulties. The recording an electroencephalogram (EEG). Regarding monitoring of fluctuating vigilance by parameters of the specific EEG correlates of RT performance, Jokeit and peripheral nervous system, such as heart rate and elec- Makeig [9] compared different EEG patterns of subjects trodermal activity (EDA), proved to be unreliable due to with quick and slow mean RT. Qualitative differences in the slow response rates of these indirect parameters. Thus, the assessment of vigilance via EEG appears to be EEG patterns were reported between these two subject an adequate approach to determine global functional groups. Examining EEG patterns of healthy subjects, Delorme et al. [10] have revealed that larger low-theta levels of the brain. However, even expert raters showed complexes precede quicker motor responses both within poor performance in identifying vigilance lapses using and across subjects. Moreover, Makeig and Jung [11] EEG [15]. Therefore, Hegerl et al. [16] developed a com- demonstrated that performance variations on an audi- puter-based algorithm (VIGALL, Vigilance Algorithm tory vigilance task show distinct EEG-correlates on dif- Leipzig) that classifies different vigilance stages of EEG ferent time-scales. segments according to Bente and Roth (A1, A2, A3, B1, The EEG is the only non-invasive method that directly B2/3), based on the frequency and topographical distri- measures neuronal activity with sufficient time resolu- bution of the neuroelectric activity. Olbrich et al. [17] tion. On the basis of specific EEG-patterns, Loomis [12], validated and refined this algorithm. Hence, VIGALL is Bente [13] and Roth [14] classified different activation now based on EEG-power source estimates using LOR- states of the brain on a continuum reaching from the ETA (Low Resolution Brain Electromagnetic Tomogra- concentrated awake state to the state of deep sleep. In phy) and enables the classification of EEG-vigilance the following, these states, which influence the ability to stages for 1-sec-segments. Figure 1 depicts decision cri- process information, are termed EEG-vigilance stages. teria of the algorithm to calculate vigilance stages from They have been carefully described and subdivided (A1, the EEG data obtained. A2, A3, B1, B2/3) depending on the frequency and topo- The goal of the present study was to determine graphic distribution of the EEG-waves (see Figure 1). whether the VIGALL-classified prestimulus state of EEG-vigilance is associated with the length of RT and may therefore explain the intra-individual variance of this dimension. We postulated that a low prestimulus EEG-vigilance state (B-stages) leads to longer RTs and that a high vigilance state (A-stages) entails shorter RTs. Additionally, we intended to conduct an explorative analysis of the relationship between the EEG-vigilance substages (A1, A2, A3, B1, B2/3) and RTs. Methods Participants To reduce the inter-subject variability to the greatest possible extent, a homogenous group of healthy female Figure 1 EEG-based definition criteria of VIGALL for vigilance students, who had undergone an extensive screening for classification according to Bente (1964) & Roth (1961). Note: Vigilance stages were sub-classified (column 2) according to Bente somatic and mental disorders, was included in this (1964) and Roth (1961). Continuous EEG-based vigilance stages from study. In total, 35 female students from 20 to 30 years full alertness to drowsiness are determined by VIGALL according to of age (M = 23.71, SD = 2.78) participated in the inves- defined decision criteria (column 1). The first column presents that tigation. These volunteers were recruited through adver- vigilance stage A is corresponding to the presence of high alpha tisement and received remuneration. All participants power. Low alpha power features vigilance stage B. VIGALL classifies substages based on EEG-power source estimates using sLORETA: A1 reported no psychiatric, neurological or serious medical (occipital ROI power (a) > = parietal and frontal ROI power(a)), A2 conditions. Physical health was screened in a semi-struc- (occipital ROI power (a) < parietal and frontal ROI power(a) and tured interview and mental health was examined accord- temporal and parietal ROI power(a) > = frontal ROI 1.5* power (a)), ing to the criteria of the diagnostic and statistical A3 (occipital ROI power (a) < parietal and frontal ROI power(a) and manual of mental disorders (DSM-IV) by applying a temporal and parietal ROI power(a) < frontal ROI 1.5* power (a)), -6 2 4 B1 (power(a+δ+θ) in one ROI = < 7.5*10 μA /mm per data German version of the Structured Clinical Interview for -6 2 point), B2/3 (power(a+δ+θ) in one ROI > 7.5*10 μA /mm4 per DSM-IV disorders (SKID-I) [18]. In order to exclude data point. The right column depicts EEG curves of native two- subjects currently abusing alcohol and drugs, general seconds-segments. alcohol and drug consumption was quantified by Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 3 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 administering the alcohol use disorders identification electrode was taped on the forehead and a reference test (AUDIT) [19] and the drug use disorders identifica- electrode was fixed on the cheek below the eye. ECG tion test (DUDIT) [20]. All subjects reported normal or electrodes were placed on the right and left wrist. The corrected-to-normal visual acuity. recordings were amplified by a 40-channel-QuickAmp A total of 11 subjects had to be excluded post-hoc unit and BrainVision 2.0 software (BrainProducts, Gilch- from the main stage analysis (EEG-vigilance stage A vs. ing, Germany), which was installed on a Microsoft Win- B). Reasons and a description of the exclusion procedure dows XP compatible computer system, was used. are specified in the data preparation section. The final Pre-processing of EEG data EEG data was pre-pro- sample comprised 24 female students with an age-range cessed with the Analyzer software package according to from 20 to 29 years (M = 23.54, SD = 2.67) for the the following steps. First, EEG raw data was filtered at EEG-vigilance main stage comparison and 5 females 70 Hz (low-pass), 0.5 Hz (high-pass) and 50 Hz (notch- from 21 to 29 years (M = 24.60, SD = 3.29) for the filter, range 5 Hz). Study-relevant EEG-segments were explorative comparison of RTs of the EEG-vigilance cut out including two 1-sec-segements before and after substages. A local ethics committee approval and writ- the relevant segments prior to target presentation. This ten informed consent from each volunteer were ensured that on- and offset effects of subsequent analy- obtained prior to the investigation according to the sis steps were avoided. Then, the independent compo- declaration of Helsinki. nent analysis (ICA)-based approach [21,22] was used for both the removal of eye artefacts and the correction of Measures and procedure EEG-channels with continuous muscle activity. After Subjects performed a 15-minute visual discrimination segmentation into consecutive one-second intervals, task. Simultaneously, an EEG was recorded in a dimmed data sets were again screened for remaining muscle, and sound attenuated room. Participants sat in a com- movement, eye and sweating artefacts. Those artefacts fortable chair in an upright position. To avoid circadian were marked for exclusion from the EEG-vigilance stage effects, all EEGs were performed in the middle of the analysis. Afterwards, complex demodulation of the EEG- afternoon. frequency bands 2-4 Hz (delta), 4-8 Hz (theta), 8-12 Hz Cognitive performance test (CPT) (alpha) and 12-25 Hz (beta) were computed for all EEG The cognitive performance task used in this investiga- channels to obtain the frequency band envelope magni- tion covered 400 randomized trials. The visual stimuli tude in μV in order to approximate the power of the consisted of bold white letters with a width of about 9 underlying signal [23]. cm and a height of about 10 cm which appeared on a Using the LORETA module of the Vision Analyzer soft- black background. The stimuli set contained the target ware, the intracortical averaged squared current densities of “X” in 70% of the trials, and the distractor “O” in 30% frequency band power in four predefined regions of inter- of the cases. Each stimulus was presented for 300 ms on ests (ROIs) were calculated. The term averaged current a computer screen in front of the sitting participants densities refers to 1) the spatial averaging of the electrical with an inter-stimulus-interval of 2000 ms. The subjects’ intracortical source estimates of each voxel included within distance to the monitor was approximately 120 cm. The the four regions of interest in occipital, parietal, temporal subjects were instructed to press a button with the and frontal cortices and 2) the temporal averaging of the index finger of their dominant hand in case of target current densities at all data points within a one-second seg- presentation. Due to the fact that the applied visual dis- ment (i.e. 100 data points for a sampling rate of 100 Hz). crimination task is very easy with only two different sti- The occipital ROI involves the occipital lobe and the muli, we expected that the rate of hits (correctly cuneus, because alpha activity during rest is most promi- detected targets) would be high while the rate of errors, nent in those areas [24]. The parietal ROI consists of the including false alarms and misses, would be low. For superior and inferior parietal lobe, where shifts of alpha this reason, our analysis focussed on the variability in power have been found during the transition phase from RT, not on precision (ER). full wakefulness to sleep [25,26]. The temporal ROI com- EEG procedure prises the inferior temporal lobe owing to most prominent EEG set-up and recording 31 electrodes (sintered sil- EEG-alpha power in the inferior lobe during light sleep ver/silver chloride) placed according to the 10-20 inter- stages [27]. The frontal ROI consists of the anterior cingu- national system with impedances kept below 10 kOhm late gyrus (ACC) and the medial frontal gyrus as the most were applied to record the EEG. Data was recorded with prominent EEG alpha power and EEG theta power during a 1 kHz sampling rate and common average was used drowsiness is located within these areas [28,29]. for reference. Additionally, an electrocardiogram (ECG) Classification of EEG vigilance stages using VIGALL and electrooculogram (EOG) were recorded to control According to EEG-source estimates in the ROIs, EEG- vigilance stages were classified by the VIGALL algorithm for cardial and ocular artefacts. For EOG-recording, one Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 4 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 (Figure 1). The used intracortical current source density concerning the number of trials of each vigilance (sub-) thresholds for stage B1 correspond to the topographical stage. According to the assumption of low ERs owing to cut-off criterion of 200 μV for Fast-Fourier transformed the simplicity of the applied CPT version, no participant EEG-data at channels F3-TP9, F4-TP10, O1-TP9 and had to be excluded because of too many faults (M = O2-TP10 as it has been used in the study by Olbrich et 1.69, SD = 1.38, range 0.36-5.40). al. [17]. Reanalysing the EEG data from this study, it VIGALL classifies 1-sec-segments of the EEG data was found that the current source densities within the prior to target presentation separately for each trial and ROIs did not exceed the reported threshold for stages subject. RTs of trials with the same vigilance classifica- tion were averaged, thus mean RTs for the different classified as stage B1. Also the reported proportions for e.g. alpha anteriorisation (stage A1-A3) correspond to EEG-vigilance (sub-) stages are available for each the topographical distribution of EEG-power that has subjects. been used in the former version of the algorithm. How- Extremely fast or slow responses were treated as miss- ever, the EEG-vigilance substages were subsumed under ing values as they potentially reflect errors such as key main stage A (A1, A2 and A3) and B (B1 and B2/3) for malfunctions or accidental keystrokes. The computation the analysis of RT differences between high and low vig- of the mean RTs comprised all trials with response ilance states. Lower vigilance stages than B2/3, charac- times between 200 ms and 1000 ms. Missing values also terised by K-complexes and sleep spindles, did not resulted from non-classifiable EEG segments owing to occur within data sets. For statistical analyses, only the artefacts. In total, between 234 and 280 (MW = 266.46, vigilance stages that occurred 1 sec prior to target pre- SD = 12.35) responses were used to compute the mean sentation were evaluated. RTs of the subjects. Data preparation Statistics Four data sets had to be excluded due to lacking quality All data were processed using the PASW Statistics 18.0 of the recorded EEGs: In two cases, the raw data con- Package for Windows. The hypothesis that vigilance tained more than twenty percent of segments with arte- influences the speed of reaction was examined by apply- facts; another two recordings had no impedance ing a paired t-test for the EEG-vigilance main stage ana- information and were excluded for this reason. lysis (A vs. B) and a variance analysis for repeated Since an unbalanced distribution of vigilance stages (e. measures for the EEG-vigilance substage analysis (A1, g. the exclusive presence of one vigilance stage) would A2, A3, B1, B2/3). Hypotheses were tested two-tailed. A make it unfeasible to test the study hypotheses, a mini- probability p value of less than 0.05 was considered sig- mum of vigilance variability within the same individual nificant, whereas marginal trends were determined up to is necessary. For this reason, a minimum of 5% of each a significance level of 0.10. Normal distribution was vigilance (sub-) stage was set as a further inclusion cri- tested by the Kolmogorov-Smirnov test. All variables terion for the comparability of RT differences. There- were normally distributed. fore, seven subjects had to be excluded for the comparison of RTs of stage A with RTs of stage B Results (EEG-vigilance main stage analysis). Hence, 24 subjects Vigilance were included in further main stage analysis. Only five The distribution of the different states of wakefulness subjects displayed at least 5% of each EEG-vigilance sub- was determined for each participant. Overall, B-stages stage(A1,A2, A3,B1, B2/3)andwere includedin the (M = 54.33%, SD = 26.11) were registered slightly more additional explorative analysis of EEG-vigilance sub- frequent than A-stages (M = 45.67%, SD = 26.11), how- stages. Figure 2 features descriptive information ever, this was not statistically significant (t(23) = -0.812, p = 0.425). Vigilance and RT The mean RT was calculated individually for all partici- pants for each of the main vigilance stages. A paired t- test was used to assess the difference between response times of the two different conditions (vigilance stage A vs.B). Thedifferencebetween the individualRTin the vigilance stages A vs. B was statistically significant (t(23) = -2.805, p < 0.05). Individual mean RTs were signifi- Figure 2 The number of single trials for each vigilance (sub-) cantly shorter for events following high EEG-vigilance stage. stage A (M = 380.60 ms, SD = 44.91 ms) compared to Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 5 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 the lower EEG-vigilance stage B (M = 388.37 ms, SD = 44.15 ms). For the individual RTs of every volunteer see Figure 3. For the explorative analysis of the relationship of EEG- vigilance substages and RT, an ANOVA for repeated measurements (5 levels: A1, A2, A3, B1, B2/3) was pro- cessed. Results (F(4;16) = 2.643) of the ANOVA pro- vided marginal significance (p = 0.072) for the main effect EEG-vigilance stage. The substages of the main EEG-vigilance stage B entailed longer RTs than the sub- stages of the EEG-vigilance stage A. Furthermore, a trend of gradually increasing RT was observable within the EEG-vigilance stage A (see Figure 4). Discussion Figure 4 Mean RT referring to the EEG-vigilance substages (n = In accordance with our hypothesis, we found that in a 5). Note: The error bars represent the standard error of the mean. continuous performance task, RT depends on the presti- mulus EEG-vigilance stage. Faster individual reaction was observed during a higher EEG-vigilance level (stage study were misinterpreted as low subvigil stages B2/3 A), whereas a declining response speed was detected for although they might reflect a higher vigilance stage. As lower states of EEG-vigilance (stage B). Furthermore, a consequence, RT of substages B2/3 show decreased marginal differences in RTs between the EEG-vigilance RTs in comparison to stage B1. Another explanation for substages were found despite the small sample size of this inconsistency of RT alteration with changing EEG- this subgroup. However, a continuous increase in RT vigilancesubstages mightbethe smallsamplesizeora with decreasing vigilance levels was only found for the lacking sensitivity of the computer-based vigilance algo- substages A1, A2 and A3. Decline of vigilance from sub- rithm under the condition of open-eyed-EEG recordings. stage B1 to B2/3 did not yield the expected increase in Previous studies assessing factors influencing RT RT. variability either primarily focused on patient samples A reason for the latter result might be that stage B2/3 [5,6] or identified variables that describe differences is defined as an EEG-vigilance stage with low alpha between subjects, for instance gender and age [2]. Thus, power but high delta and theta power. Especially, the influence of transient within-subject factors on cog- increased phasic theta power has been associated with nitive performance tests, such as fluctuations in motiva- cognitive performance during cognitive tasks [30]. In tion or wakefulness, has not been considered adequately contrast to this, frontal theta power also increased dur- in previous studies. Nevertheless, an early examination ing rest without mental occupation as a sign of a further by Lansing et al. [32] described the influence of alert- decline of vigilance [31]. VIGALL originally was ness, determined by patterns of alpha rhythm in EEG, intended for classification of EEG-vigilance stages during on RT. The authors showed that subjects displayed fas- rest and hence it is possible that stages with high theta ter RTs in the alerted than in the non-alerted condition. power during the cognitive performance test within this These results are consistent with our findings of shorter RTs in case of high-vigilant states. However, the experi- ment deviated from our methods as alertness was induced by alarm signals. We determined the vigilance state without exerting an influence on vigilance shifts. Moreover, in our study the classification of vigilance states was computed automatically by the EEG-based algorithm VIGALL, whereas certain EEG-patterns in the previous study were detected by individual raters. Hence, our method to classify vigilance is certainly more economic and reliable and might therefore be broadly applicable in future measurements. Also, in a study on the coherence of fluctuations in Figure 3 Individual RT (n = 24) referring to the main EEG- performance and EEG spectrum, Makeig and Inlow [33] vigilance stages A and B. Note: The error bars represent the reported highly positive correlations between EEG individual standard deviations. power below 6-7 Hz, error rate and highly negative Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 6 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 correlations near 10 Hz. The finding of increased ERs homogeneous group of healthy female students was during the appearance of delta and theta waves in the included in the study. It remains uncertain whether the EEG corroborate our results that performance becomes findings can be transferred to other cohorts. Investiga- poorer during lower states of vigilance. The observations tions with different samples are required to validate the by Makeig and Inlow are in agreement with Jung et al. observed influence of vigilance states on RTs. In addi- [34], who reported a correlation between increased ER tion, the sample size of 24 healthy participants was small. This statistical drawback becomes more notice- and EEG power below 5 Hz. able for the analysis of the vigilance substages, as the Hultsch and colleagues [5] suggested that intra-indivi- sample size for the explanatory ANOVA analysis is very dual variability might be a marker for impaired neurolo- gical functioning, as patients with mild dementia small with only 5 subjects included. The vigilance effect showed twice as much intra-individual variability in per- should be validated by future studies with larger num- formance as neurologically intact participants. There- bers of patients and healthy controls. Furthermore, as fore, monitoring of factors which cause within-subject only an easy visual discrimination task was carried out, variations during a performance task is essential to eval- the obtained findings can not be generalized for all RT uate the observed individual variability. As demonstrated paradigms. According to Stuss et al. [35], RT variability in the present study, the covariate vigilance explains the might increase in more complex tasks. Therefore, sev- variance to a certain extent. Vigilance-specific alterations eral RT tasks involving different qualities of sensory per- of electric brain activity during performance tasks ception and various levels of difficulty should be should thus be taken into account. By disregarding indi- implemented. Another drawback of the study is that the vidual vigilance-driven electrical brain signal fluctua- used version of VIGALL could not be adjusted individu- tions, relations between intra-individual variability of ally, as it used equal EEG criteria for all participants. RTs and neurological disease might be either over- or Currently, the VIGALL team is refining the operating underestimated. mode of the algorithm taking into account the indivi- Hence, the present study is methodologically impor- dual alpha peak instead of fixed frequency windows for tant by emphasising the necessity of considering vigi- VIGALL processing. Consequently, individual bound- lance in studies whilst planning, performing and aries of frequency bands could be justified, making the interpreting cognitive tasks. We determined vigilance vigilance classification of VIGALL more exact. using EEG, a well-established and non-invasive method Of course, vigilance is only one possible factor which might influence RT in real life. For example, age [36], in medicine. By this method, vigilance monitoring and alcohol [37], drugs [38], certain psychiatric [39] and classification is broadly applicable in future studies to control for intra-individual variability. somatic diseases [40] as well as distraction [41] have Furthermore, the observation that vigilance affects been shown to influence RT. Our results suggest that behavioural measures opens up perspectives to further hormones which influence the sleep-wake regulation improve the validity of neuro-imaging methods. For and therefore vigilance such as glucocorticoids, melato- instance, functional imaging studies might profit from nin and leptin as well as other hormones which influ- eliminating unexplained intra-individual variance by tak- ence these endocrine systems such as estrogens, ing different states of vigilance into account. Therefore, androgens and thyroid hormones might also play a role simultaneous usage of functional imaging methods and as influencing factor on reaction time. Therefore, we are EEG is beneficial. Technical requirements have already going to investigate the influence of these hormones on been met and PET- and fMRI-compatible EEG instru- RT and differences regarding influencing factors on RT ments are now available. Thus, the covariate vigilance in females and males in future studies. Due to fact that can be controlled for by monitoring for vigilance shifts this is a pilot study using a small group of homogenous using the VIGALL algorithm during neuro-imaging pro- healthy female subjects, we have to discuss the limita- cedures. When analysing and interpreting neuro-ima- tion that we are not able to give any data regarding ging data sets, EEG-vigilance stages could either be these mentioned additional possible influencing factors considered in the general evaluation, or only those time such as age, gender, alcohol, drugs, psychiatric and segments could be taken into account that show certain somatic diseases, distraction and hormones. Another vigilance states. Further studies are needed to assess the important issue to consider is specific methodological general impact of vigilance states on neuro-imaging approach of this study. The applied CPT was very easy and monotonous, since we intended to induce a vigi- methods. lance decline. The inter-stimulus-interval was 2000 ms. Despite these advantages of the presented study, some Furthermore, the smallest possible VIGALL analysis unit limitations also have to be mentioned. One shortcoming of our examination is the fact that, for reasons of mini- is one second. We decided to analyze the whole second mising the inter-individual variability, a rather before target presentation and not while or after target Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 7 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 8. Oken BS, Salinsky MC, Elsas SM: Vigilance, alertness, or sustained presentation, because a) the stimulus presentation attention: physiological basis and measurement. Clin Neurophysiol 2006, induces an arousal and modifies the EEG and b) the 117:1885-1901. subjects’ reaction leads to artefacts in the EEG and is 9. Jokeit H, Makeig S: Different event-related patterns of γ-band power in brain waves of fast and slow-reacting subjects. Proc Natl Acad Sci USA expected 300-500 ms after target presentation. There- 1994, 91:6339-6343. fore, our results regarding vigilance are-with good rea- 10. Delorme A, Westerfield M, Makeig S: Medial prefrontal theta bursts son-limited to the one second before target presentation. precede rapid motor responses during visual selective attention. J Neurosci 2007, 27:11949-11959. In summary, the results of the present study demon- 11. Makeig S, Jung T-P: Tonic, phasic, and transient EEG correlates of strate that RT is associated with prestimulus vigilance as auditory awareness in drowsiness. Brain Res Cogn Brain Res 1996, 4:15-25. it can be measured using EEG. By including this EEG- 12. Davis H, Davis PA, Loomis AL, Harvey EN, Hobart G: Changes in human brain potentials during the onset of sleep. Science 1937, 86:448-450. measured vigilance in the analysis, the conclusiveness of 13. Bente D: Vigilanz, dissoziative Vigilanzverschiebung und Insuffizienz des scientific data on cognitive performance or reaction tasks Vigilitätstonus. In Begleitwirkung und Mißerfolge der psychiatrischen might be improved. This may be relevant for neuropsy- Pharmakotherapie. Edited by: Kranz H, Heinrich K. Stuttgart: Thieme; 1964:13-28. chological as well as for functional neuroimaging studies. 14. Roth B: Clinical and theoretical importance of EEG rhythms corresponding to states of lowered vigilance. Electroencephalogr Clin Neurophysiol 1961, 13:395-399. Acknowledgements 15. Peiris MT, Jones RD, Davidson PR, Carroll GJ, Signal TL, Parkin PJ, van den Publication of this study was supported by the Claussen-Simon-Foundation. Berg M, Bones PJ: Identification of vigilance lapses using EEG/EOG by expert human raters. Proceedings of the Annual International Conference of Author details the IEEE Engineering in Medicine and Biology Society: 1-4 September 2005; Department of Psychiatry and Psychotherapy, University of Leipzig, Shanghai 2006, 6:5735-5737. Semmelweisstr. 10, 04103 Leipzig, Germany. Department of Psychiatry and 16. Hegerl U, Stein M, Mulert C, Mergl R, Olbrich S, Dichgans E, Rujescu D, Psychotherapy, Medical Faculty, RWTH Aachen University, Pauwelsstr. 30, Pogarell O: EEG-vigilance differences between patients with borderline 52074 Aachen, Germany. personality disorder, patients with obsessive-compulsive disorder and healthy controls. Eur Arch of Psychiatry Clin Neurosci 2005, 258:137-143. Authors’ contributions 17. Olbrich S, Sander C, Jahn I, Eplinius F, Claus S, Mergl R, Schönknecht P, The presented work was carried out in collaboration between all authors. All Hegerl U: Unstable EEG-vigilance in patients with cancer-related fatique authors were involved in drafting and revising critically the manuscript and in comparison to healthy controls. World J Biol Psychiatry . approved the final version for publication. PS and UH defined the research 18. Wittchen H-U, Zaudig M, Fydrich T: SKID-Strukturiertes Klinisches theme and planned the conception of the study. MT and CS designed the Interview für DSM-IV. Achse I und II Göttingen: Hogrefe; 1996. experiment methods and acquired data. JM and HH made substantial 19. Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro MG: The Alcohol Use contributions to the conception and design, analyzed the data, interpreted Disorders Identification Test: Guidelines for use in primary care. Geneva: the results and wrote the paper. SO and AS directed the EEG recording WHO;, 2 1989. methods and discussed analyses, interpretation and presentation. 20. Berman AH, Bergman H, Palmstierna T, Schlyter F: Evaluation of the Drug Use Disorders Identification Test (DUDIT) in criminal justice and Competing interests detoxification settings and in a Swedish population sample. Eur Addict All authors declare not to have any conflict of interest including any Res 2005, 11:22-31. financial, personal or other relationships with other people or organizations 21. Delorme A, Sejnowski T, Makeig S: Enhanced detection of artefacts in EEG that could inappropriately influence, or be perceived to influence, their work. data using higher-order statistics and independent component analysis. Neuroimage 2007, 34:1443-1449. Received: 21 April 2011 Accepted: 5 August 2011 22. Olbrich S, Jödicke J, Sander C, Himmerich H, Hegerl U: ICA-based muscle Published: 5 August 2011 artefact correction of EEG data: What is muscle and what is brain? Comment on McMenamin et al. Neuroimage 2011, 54:4-9. References 23. Schroeder MJ, Barr RE: An alpha modulation index for 1. MacDonald SW, Hultsch DF, Bunce D: Intra-individual variability in electroencephalographic studies using complex demodulation. Med Biol vigilance performance: does degrading visual stimuli mimic age-related Eng Comput 2000, 38:306-310. “neural noise"? J Clin Exp Neuropsycho 2006, 28:655-675. 24. Niedermeyer E: Alpha rhythms as physiological and abnormal 2. Reimers S, Maylor EA: Gender effects on reaction time variability and phenomena. Int J Psychophysiol 1997, 26:31-49. trial-to-trial performance: reply to Deary and Der (2005). Neuropsychol 25. De Gennaro L, Ferrara M, Curcio G, Cristiani R: Antero-posterior EEG Dev Cogn B Aging Neuropsychol Cogn 2006, 13:479-489. changes during the wakefulness-sleep transition. Clin Neurophysiol 2001, 3. Myerson J, Robertson S, Hale S: Aging and intra-individual variability in 112:1901-1911. performance: analyses of response time distributions. J Exp Anal Behav 26. Manshanden I, De Munck JC, Simon NR, Lopes da Silva FH: Source 2007, 88:319-337. localization of MEG sleep spindles and the relation to sources of alpha 4. MacDonald SW, Nyberg L, Sandblom J, Fischer H, Bäckman L: Increased band rhythms. Clin Neurophysiol 2002, 113:1937-1947. response-time variability is associated with reduced inferior parietal 27. Coatanhay A, Soufflet L, Staner L, Boeijinga P: EEG source identification: activation during episodic recognition in aging. J Cogn Neurosci 2008, frequency analysis during sleep. C R Biol 2002, 325:273-282. 20:779-786. 28. Connemann BJ, Mann K, Lange-Asschenfeldt C, Ruchsow M, 5. Hultsch DF, MacDonald SW, Hunter MA, Levy-Bencheton J, Strauss E: Intra- Schreckenberger M, Bartenstein P, Gründer G: Anterior limbic alpha-like individual variability in cognitive performance in older adults: activity: a low resolution electromagnetic tomography study with comparison of adults with mild dementia, adults with arthritis, and lorazepam challenge. Clin Neurophysiol 2005, 116:886-894. healthy adults. Neuropsychology 2000, 14:588-598. 29. Broughton R, Hasan J: Quantitative topographic electroencephalographic 6. De Frias CM, Dixon RA, Fisher N, Camicioli R: Intra-individual variability in mapping during drowsiness and sleep onset. J Clin Neurophysiol 1995, neurocognitive speed: a comparison of Parkinson’s disease and normal 12:372-386. older adults. Neuropsychologia 2007, 45:2499-2507. 30. Klimesch W: EEG alpha and theta oscillations reflect cognitive and 7. Jensen AR: The importance of intra-individual variation in reaction time. memory performance: a review and analysis. Brain Res Brain Res Rev 1999, Pers Individ Dif 1992, 13:869-881. 29:169-195. Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 8 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 31. Strijkstra AM, Beersma DG, Drayer B, Halbesma N, Daan S: Subjective sleepiness correlates negatively with global alpha (8-12 Hz) and positively with central frontal theta (4-8 Hz) frequencies in the human resting awake electroencephalogram. Neurosci Lett 2003, 340:17-20. 32. Lansing RW, Schwartz E, Lindsley DB: Reaction time and EEG activation under alerted and nonalerted conditions. J Exp Psychol 1959, 58:1-7. 33. Makeig S, Inlow M: Lapses in alertness: coherence of fluctuations in performance and EEG spectrum. Electroencephalogr Clin Neurophysiol 1993, 86:23-35. 34. Jung TP, Makeig S, Stensmo M, Sejnowski TJ: Estimating alertness from the EEG power spectrum. IEEE Trans Biomed Eng 1997, 44:60-69. 35. Stuss DT, Pogue J, Buckle L, Bondar J: Characterization of stability of performance in patients with traumatic brain injury: variability and consistency on reaction time tests. Neuropsychology 1994, 8:316-324. 36. Anstey KJ, Dear K, Christensen H, Jorm AF: Biomarkers, health, lifestyle and demographic variables as correlates of reaction time performance in early, middle, and late adulthood. Q J Exp Psychol 2005, 58:5-21. 37. Howland J, Rohsenow DJ, Arnedt JT, Bliss CA, Hunt SK, Calise TV, Heeren T, Winter M, Littlefield C, Gottlieb DJ: The acute effects of caffeinated versus non-caffeinated alcoholic beverage on driving performance and attention/reaction time. Addiction 2011, 106:335-341. 38. Brockel BJ, Fowler SC: Effects of chronic haloperidol on reaction time and errors in a sustained attention task: partial reversal by anticholinergics and by amphetamine. J Pharmacol Exp Ther 1995, 275:1090-1098. 39. Kaiser S, Roth A, Rentrop M, Friederich H-C, Bender S, Weisbrod M: Intra- individual reaction time variability in schizophrenia, depression and borderline personality disorder. Brain Cogn 2008, 66:73-82. 40. Mellgren A, Norkrans G, Hagberg L, Dunlop O, Wejstal R, Gisslen M: Slowed reaction time in HIV-1-infected patients without AIDS. Acta Neurol Scand 2000, 102:169-174. 41. Travis F, Tecce JJ: Effects of distracting stimuli on CNV amplitude and reaction time. Int J Psychophysiol 1998, 31:45-50. doi:10.1186/1744-9081-7-31 Cite this article as: Minkwitz et al.: Prestimulus vigilance predicts response speed in an easy visual discrimination task. Behavioral and Brain Functions 2011 7:31. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral and Brain Functions Springer Journals

Prestimulus vigilance predicts response speed in an easy visual discrimination task

Download PDF
8 pages

Loading next page...
 
/lp/springer-journals/prestimulus-vigilance-predicts-response-speed-in-an-easy-visual-wJrkTNtvnc

References (51)

Publisher
Springer Journals
Copyright
Copyright © 2011 by Minkwitz et al; licensee BioMed Central Ltd.
Subject
Biomedicine; Neurosciences; Neurology; Behavioral Therapy; Psychiatry
eISSN
1744-9081
DOI
10.1186/1744-9081-7-31
pmid
21816115
Publisher site
See Article on Publisher Site

Abstract

Background: Healthy adults show considerable within-subject variation of reaction time (RT) when performing cognitive tests. So far, the neurophysiological correlates of these inconsistencies have not yet been investigated sufficiently. In particular, studies rarely have focused on alterations of prestimulus EEG-vigilance as a factor which possibly influences the outcome of cognitive tests. We hypothesised that a low EEG-vigilance state immediately before a reaction task would entail a longer RT. Shorter RTs were expected for a high EEG-vigilance state. Methods: 24 female students performed an easy visual discrimination task while an electroencephalogram (EEG) was recorded. The vigilance stages of 1-sec-EEG-segments before stimulus presentation were classified automatically using the computer-based Vigilance Algorithm Leipzig (VIGALL). The mean RTs of each EEG-vigilance stage were calculated for each subject. A paired t-test for the EEG-vigilance main stage analysis (A vs. B) and a variance analysis for repeated measures for the EEG-vigilance sub-stage analysis (A1, A2, A3, B1, B2/3) were calculated. Results: Individual mean RT was significantly shorter for events following the high EEG-vigilance stage A compared to the lower EEG-vigilance stage B. The main effect of the sub-stage analysis was marginal significant. A trend of gradually increasing RT was observable within the EEG-vigilance stage A. Conclusion: We conclude that an automatically classified low EEG-vigilance level is associated with an increased RT. Thus, intra-individual variances in cognitive test might be explainable in parts by the individual state of EEG- vigilance. Therefore, the accuracy of neuro-cognitive investigations might be improvable by simultaneously controlling for vigilance shifts using the EEG and VIGALL. Introduction High intra-individual variance of behavioural measures Typically, studies in cognitive neuroscience implement leads to biased evaluations of cognitive processing, both paradigms, e.g. cognitive performance tasks, in which within healthy subjects as well as patient cohorts. Pro- participants respond to randomly presented sensory sti- longation of an experimental paradigm is a frequently muli. By comparing averages of stimulus-locked used and common method to reduce intra-individual responses, such as reaction time (RT) or error rate (ER), variance for obtaining a more reliable measure of RT valuable information on cognitive processing can be [7]. Nevertheless, the amount of experimental trials can gained. Beyond the variability between different subjects be limited, e.g. due to reduced physical and mental con- dition of patients or older subjects. Thus, intra-indivi- (inter-individual variability), responses of the same sub- ject vary crucially (intra-individual variability) across dual variability should be minimised by controlling for experiments [1]. Previous studies report that inter-indi- potential covariates that directly influence RT and ER. vidual differences in RT are associated with gender, age Therefore, we focused on the fluctuating state of [2] and neurological alterations [3-6]. wakefulness, as we hypothesized that the level of alert- ness crucially impacts individual performance during an * Correspondence: Hubertus.Himmerich@medizin.uni-leipzig.de experimental procedure. For examining our hypothesis, Department of Psychiatry and Psychotherapy, University of Leipzig, we adhere to the EEG-vigilance concept, which unfortu- Semmelweisstr. 10, 04103 Leipzig, Germany nately overlaps with other concepts, e.g. alertness, Full list of author information is available at the end of the article © 2011 Minkwitz et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 2 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 attention and arousal [8]. We use the term vigilance to Nevertheless, examination of EEG-vigilance based refer to different levels of brain function on the sleep- variability in RT tasks has not been used on a single- wake spectrum as they are empirically assessable by trial basis so far due to methodological difficulties. The recording an electroencephalogram (EEG). Regarding monitoring of fluctuating vigilance by parameters of the specific EEG correlates of RT performance, Jokeit and peripheral nervous system, such as heart rate and elec- Makeig [9] compared different EEG patterns of subjects trodermal activity (EDA), proved to be unreliable due to with quick and slow mean RT. Qualitative differences in the slow response rates of these indirect parameters. Thus, the assessment of vigilance via EEG appears to be EEG patterns were reported between these two subject an adequate approach to determine global functional groups. Examining EEG patterns of healthy subjects, Delorme et al. [10] have revealed that larger low-theta levels of the brain. However, even expert raters showed complexes precede quicker motor responses both within poor performance in identifying vigilance lapses using and across subjects. Moreover, Makeig and Jung [11] EEG [15]. Therefore, Hegerl et al. [16] developed a com- demonstrated that performance variations on an audi- puter-based algorithm (VIGALL, Vigilance Algorithm tory vigilance task show distinct EEG-correlates on dif- Leipzig) that classifies different vigilance stages of EEG ferent time-scales. segments according to Bente and Roth (A1, A2, A3, B1, The EEG is the only non-invasive method that directly B2/3), based on the frequency and topographical distri- measures neuronal activity with sufficient time resolu- bution of the neuroelectric activity. Olbrich et al. [17] tion. On the basis of specific EEG-patterns, Loomis [12], validated and refined this algorithm. Hence, VIGALL is Bente [13] and Roth [14] classified different activation now based on EEG-power source estimates using LOR- states of the brain on a continuum reaching from the ETA (Low Resolution Brain Electromagnetic Tomogra- concentrated awake state to the state of deep sleep. In phy) and enables the classification of EEG-vigilance the following, these states, which influence the ability to stages for 1-sec-segments. Figure 1 depicts decision cri- process information, are termed EEG-vigilance stages. teria of the algorithm to calculate vigilance stages from They have been carefully described and subdivided (A1, the EEG data obtained. A2, A3, B1, B2/3) depending on the frequency and topo- The goal of the present study was to determine graphic distribution of the EEG-waves (see Figure 1). whether the VIGALL-classified prestimulus state of EEG-vigilance is associated with the length of RT and may therefore explain the intra-individual variance of this dimension. We postulated that a low prestimulus EEG-vigilance state (B-stages) leads to longer RTs and that a high vigilance state (A-stages) entails shorter RTs. Additionally, we intended to conduct an explorative analysis of the relationship between the EEG-vigilance substages (A1, A2, A3, B1, B2/3) and RTs. Methods Participants To reduce the inter-subject variability to the greatest possible extent, a homogenous group of healthy female Figure 1 EEG-based definition criteria of VIGALL for vigilance students, who had undergone an extensive screening for classification according to Bente (1964) & Roth (1961). Note: Vigilance stages were sub-classified (column 2) according to Bente somatic and mental disorders, was included in this (1964) and Roth (1961). Continuous EEG-based vigilance stages from study. In total, 35 female students from 20 to 30 years full alertness to drowsiness are determined by VIGALL according to of age (M = 23.71, SD = 2.78) participated in the inves- defined decision criteria (column 1). The first column presents that tigation. These volunteers were recruited through adver- vigilance stage A is corresponding to the presence of high alpha tisement and received remuneration. All participants power. Low alpha power features vigilance stage B. VIGALL classifies substages based on EEG-power source estimates using sLORETA: A1 reported no psychiatric, neurological or serious medical (occipital ROI power (a) > = parietal and frontal ROI power(a)), A2 conditions. Physical health was screened in a semi-struc- (occipital ROI power (a) < parietal and frontal ROI power(a) and tured interview and mental health was examined accord- temporal and parietal ROI power(a) > = frontal ROI 1.5* power (a)), ing to the criteria of the diagnostic and statistical A3 (occipital ROI power (a) < parietal and frontal ROI power(a) and manual of mental disorders (DSM-IV) by applying a temporal and parietal ROI power(a) < frontal ROI 1.5* power (a)), -6 2 4 B1 (power(a+δ+θ) in one ROI = < 7.5*10 μA /mm per data German version of the Structured Clinical Interview for -6 2 point), B2/3 (power(a+δ+θ) in one ROI > 7.5*10 μA /mm4 per DSM-IV disorders (SKID-I) [18]. In order to exclude data point. The right column depicts EEG curves of native two- subjects currently abusing alcohol and drugs, general seconds-segments. alcohol and drug consumption was quantified by Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 3 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 administering the alcohol use disorders identification electrode was taped on the forehead and a reference test (AUDIT) [19] and the drug use disorders identifica- electrode was fixed on the cheek below the eye. ECG tion test (DUDIT) [20]. All subjects reported normal or electrodes were placed on the right and left wrist. The corrected-to-normal visual acuity. recordings were amplified by a 40-channel-QuickAmp A total of 11 subjects had to be excluded post-hoc unit and BrainVision 2.0 software (BrainProducts, Gilch- from the main stage analysis (EEG-vigilance stage A vs. ing, Germany), which was installed on a Microsoft Win- B). Reasons and a description of the exclusion procedure dows XP compatible computer system, was used. are specified in the data preparation section. The final Pre-processing of EEG data EEG data was pre-pro- sample comprised 24 female students with an age-range cessed with the Analyzer software package according to from 20 to 29 years (M = 23.54, SD = 2.67) for the the following steps. First, EEG raw data was filtered at EEG-vigilance main stage comparison and 5 females 70 Hz (low-pass), 0.5 Hz (high-pass) and 50 Hz (notch- from 21 to 29 years (M = 24.60, SD = 3.29) for the filter, range 5 Hz). Study-relevant EEG-segments were explorative comparison of RTs of the EEG-vigilance cut out including two 1-sec-segements before and after substages. A local ethics committee approval and writ- the relevant segments prior to target presentation. This ten informed consent from each volunteer were ensured that on- and offset effects of subsequent analy- obtained prior to the investigation according to the sis steps were avoided. Then, the independent compo- declaration of Helsinki. nent analysis (ICA)-based approach [21,22] was used for both the removal of eye artefacts and the correction of Measures and procedure EEG-channels with continuous muscle activity. After Subjects performed a 15-minute visual discrimination segmentation into consecutive one-second intervals, task. Simultaneously, an EEG was recorded in a dimmed data sets were again screened for remaining muscle, and sound attenuated room. Participants sat in a com- movement, eye and sweating artefacts. Those artefacts fortable chair in an upright position. To avoid circadian were marked for exclusion from the EEG-vigilance stage effects, all EEGs were performed in the middle of the analysis. Afterwards, complex demodulation of the EEG- afternoon. frequency bands 2-4 Hz (delta), 4-8 Hz (theta), 8-12 Hz Cognitive performance test (CPT) (alpha) and 12-25 Hz (beta) were computed for all EEG The cognitive performance task used in this investiga- channels to obtain the frequency band envelope magni- tion covered 400 randomized trials. The visual stimuli tude in μV in order to approximate the power of the consisted of bold white letters with a width of about 9 underlying signal [23]. cm and a height of about 10 cm which appeared on a Using the LORETA module of the Vision Analyzer soft- black background. The stimuli set contained the target ware, the intracortical averaged squared current densities of “X” in 70% of the trials, and the distractor “O” in 30% frequency band power in four predefined regions of inter- of the cases. Each stimulus was presented for 300 ms on ests (ROIs) were calculated. The term averaged current a computer screen in front of the sitting participants densities refers to 1) the spatial averaging of the electrical with an inter-stimulus-interval of 2000 ms. The subjects’ intracortical source estimates of each voxel included within distance to the monitor was approximately 120 cm. The the four regions of interest in occipital, parietal, temporal subjects were instructed to press a button with the and frontal cortices and 2) the temporal averaging of the index finger of their dominant hand in case of target current densities at all data points within a one-second seg- presentation. Due to the fact that the applied visual dis- ment (i.e. 100 data points for a sampling rate of 100 Hz). crimination task is very easy with only two different sti- The occipital ROI involves the occipital lobe and the muli, we expected that the rate of hits (correctly cuneus, because alpha activity during rest is most promi- detected targets) would be high while the rate of errors, nent in those areas [24]. The parietal ROI consists of the including false alarms and misses, would be low. For superior and inferior parietal lobe, where shifts of alpha this reason, our analysis focussed on the variability in power have been found during the transition phase from RT, not on precision (ER). full wakefulness to sleep [25,26]. The temporal ROI com- EEG procedure prises the inferior temporal lobe owing to most prominent EEG set-up and recording 31 electrodes (sintered sil- EEG-alpha power in the inferior lobe during light sleep ver/silver chloride) placed according to the 10-20 inter- stages [27]. The frontal ROI consists of the anterior cingu- national system with impedances kept below 10 kOhm late gyrus (ACC) and the medial frontal gyrus as the most were applied to record the EEG. Data was recorded with prominent EEG alpha power and EEG theta power during a 1 kHz sampling rate and common average was used drowsiness is located within these areas [28,29]. for reference. Additionally, an electrocardiogram (ECG) Classification of EEG vigilance stages using VIGALL and electrooculogram (EOG) were recorded to control According to EEG-source estimates in the ROIs, EEG- vigilance stages were classified by the VIGALL algorithm for cardial and ocular artefacts. For EOG-recording, one Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 4 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 (Figure 1). The used intracortical current source density concerning the number of trials of each vigilance (sub-) thresholds for stage B1 correspond to the topographical stage. According to the assumption of low ERs owing to cut-off criterion of 200 μV for Fast-Fourier transformed the simplicity of the applied CPT version, no participant EEG-data at channels F3-TP9, F4-TP10, O1-TP9 and had to be excluded because of too many faults (M = O2-TP10 as it has been used in the study by Olbrich et 1.69, SD = 1.38, range 0.36-5.40). al. [17]. Reanalysing the EEG data from this study, it VIGALL classifies 1-sec-segments of the EEG data was found that the current source densities within the prior to target presentation separately for each trial and ROIs did not exceed the reported threshold for stages subject. RTs of trials with the same vigilance classifica- tion were averaged, thus mean RTs for the different classified as stage B1. Also the reported proportions for e.g. alpha anteriorisation (stage A1-A3) correspond to EEG-vigilance (sub-) stages are available for each the topographical distribution of EEG-power that has subjects. been used in the former version of the algorithm. How- Extremely fast or slow responses were treated as miss- ever, the EEG-vigilance substages were subsumed under ing values as they potentially reflect errors such as key main stage A (A1, A2 and A3) and B (B1 and B2/3) for malfunctions or accidental keystrokes. The computation the analysis of RT differences between high and low vig- of the mean RTs comprised all trials with response ilance states. Lower vigilance stages than B2/3, charac- times between 200 ms and 1000 ms. Missing values also terised by K-complexes and sleep spindles, did not resulted from non-classifiable EEG segments owing to occur within data sets. For statistical analyses, only the artefacts. In total, between 234 and 280 (MW = 266.46, vigilance stages that occurred 1 sec prior to target pre- SD = 12.35) responses were used to compute the mean sentation were evaluated. RTs of the subjects. Data preparation Statistics Four data sets had to be excluded due to lacking quality All data were processed using the PASW Statistics 18.0 of the recorded EEGs: In two cases, the raw data con- Package for Windows. The hypothesis that vigilance tained more than twenty percent of segments with arte- influences the speed of reaction was examined by apply- facts; another two recordings had no impedance ing a paired t-test for the EEG-vigilance main stage ana- information and were excluded for this reason. lysis (A vs. B) and a variance analysis for repeated Since an unbalanced distribution of vigilance stages (e. measures for the EEG-vigilance substage analysis (A1, g. the exclusive presence of one vigilance stage) would A2, A3, B1, B2/3). Hypotheses were tested two-tailed. A make it unfeasible to test the study hypotheses, a mini- probability p value of less than 0.05 was considered sig- mum of vigilance variability within the same individual nificant, whereas marginal trends were determined up to is necessary. For this reason, a minimum of 5% of each a significance level of 0.10. Normal distribution was vigilance (sub-) stage was set as a further inclusion cri- tested by the Kolmogorov-Smirnov test. All variables terion for the comparability of RT differences. There- were normally distributed. fore, seven subjects had to be excluded for the comparison of RTs of stage A with RTs of stage B Results (EEG-vigilance main stage analysis). Hence, 24 subjects Vigilance were included in further main stage analysis. Only five The distribution of the different states of wakefulness subjects displayed at least 5% of each EEG-vigilance sub- was determined for each participant. Overall, B-stages stage(A1,A2, A3,B1, B2/3)andwere includedin the (M = 54.33%, SD = 26.11) were registered slightly more additional explorative analysis of EEG-vigilance sub- frequent than A-stages (M = 45.67%, SD = 26.11), how- stages. Figure 2 features descriptive information ever, this was not statistically significant (t(23) = -0.812, p = 0.425). Vigilance and RT The mean RT was calculated individually for all partici- pants for each of the main vigilance stages. A paired t- test was used to assess the difference between response times of the two different conditions (vigilance stage A vs.B). Thedifferencebetween the individualRTin the vigilance stages A vs. B was statistically significant (t(23) = -2.805, p < 0.05). Individual mean RTs were signifi- Figure 2 The number of single trials for each vigilance (sub-) cantly shorter for events following high EEG-vigilance stage. stage A (M = 380.60 ms, SD = 44.91 ms) compared to Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 5 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 the lower EEG-vigilance stage B (M = 388.37 ms, SD = 44.15 ms). For the individual RTs of every volunteer see Figure 3. For the explorative analysis of the relationship of EEG- vigilance substages and RT, an ANOVA for repeated measurements (5 levels: A1, A2, A3, B1, B2/3) was pro- cessed. Results (F(4;16) = 2.643) of the ANOVA pro- vided marginal significance (p = 0.072) for the main effect EEG-vigilance stage. The substages of the main EEG-vigilance stage B entailed longer RTs than the sub- stages of the EEG-vigilance stage A. Furthermore, a trend of gradually increasing RT was observable within the EEG-vigilance stage A (see Figure 4). Discussion Figure 4 Mean RT referring to the EEG-vigilance substages (n = In accordance with our hypothesis, we found that in a 5). Note: The error bars represent the standard error of the mean. continuous performance task, RT depends on the presti- mulus EEG-vigilance stage. Faster individual reaction was observed during a higher EEG-vigilance level (stage study were misinterpreted as low subvigil stages B2/3 A), whereas a declining response speed was detected for although they might reflect a higher vigilance stage. As lower states of EEG-vigilance (stage B). Furthermore, a consequence, RT of substages B2/3 show decreased marginal differences in RTs between the EEG-vigilance RTs in comparison to stage B1. Another explanation for substages were found despite the small sample size of this inconsistency of RT alteration with changing EEG- this subgroup. However, a continuous increase in RT vigilancesubstages mightbethe smallsamplesizeora with decreasing vigilance levels was only found for the lacking sensitivity of the computer-based vigilance algo- substages A1, A2 and A3. Decline of vigilance from sub- rithm under the condition of open-eyed-EEG recordings. stage B1 to B2/3 did not yield the expected increase in Previous studies assessing factors influencing RT RT. variability either primarily focused on patient samples A reason for the latter result might be that stage B2/3 [5,6] or identified variables that describe differences is defined as an EEG-vigilance stage with low alpha between subjects, for instance gender and age [2]. Thus, power but high delta and theta power. Especially, the influence of transient within-subject factors on cog- increased phasic theta power has been associated with nitive performance tests, such as fluctuations in motiva- cognitive performance during cognitive tasks [30]. In tion or wakefulness, has not been considered adequately contrast to this, frontal theta power also increased dur- in previous studies. Nevertheless, an early examination ing rest without mental occupation as a sign of a further by Lansing et al. [32] described the influence of alert- decline of vigilance [31]. VIGALL originally was ness, determined by patterns of alpha rhythm in EEG, intended for classification of EEG-vigilance stages during on RT. The authors showed that subjects displayed fas- rest and hence it is possible that stages with high theta ter RTs in the alerted than in the non-alerted condition. power during the cognitive performance test within this These results are consistent with our findings of shorter RTs in case of high-vigilant states. However, the experi- ment deviated from our methods as alertness was induced by alarm signals. We determined the vigilance state without exerting an influence on vigilance shifts. Moreover, in our study the classification of vigilance states was computed automatically by the EEG-based algorithm VIGALL, whereas certain EEG-patterns in the previous study were detected by individual raters. Hence, our method to classify vigilance is certainly more economic and reliable and might therefore be broadly applicable in future measurements. Also, in a study on the coherence of fluctuations in Figure 3 Individual RT (n = 24) referring to the main EEG- performance and EEG spectrum, Makeig and Inlow [33] vigilance stages A and B. Note: The error bars represent the reported highly positive correlations between EEG individual standard deviations. power below 6-7 Hz, error rate and highly negative Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 6 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 correlations near 10 Hz. The finding of increased ERs homogeneous group of healthy female students was during the appearance of delta and theta waves in the included in the study. It remains uncertain whether the EEG corroborate our results that performance becomes findings can be transferred to other cohorts. Investiga- poorer during lower states of vigilance. The observations tions with different samples are required to validate the by Makeig and Inlow are in agreement with Jung et al. observed influence of vigilance states on RTs. In addi- [34], who reported a correlation between increased ER tion, the sample size of 24 healthy participants was small. This statistical drawback becomes more notice- and EEG power below 5 Hz. able for the analysis of the vigilance substages, as the Hultsch and colleagues [5] suggested that intra-indivi- sample size for the explanatory ANOVA analysis is very dual variability might be a marker for impaired neurolo- gical functioning, as patients with mild dementia small with only 5 subjects included. The vigilance effect showed twice as much intra-individual variability in per- should be validated by future studies with larger num- formance as neurologically intact participants. There- bers of patients and healthy controls. Furthermore, as fore, monitoring of factors which cause within-subject only an easy visual discrimination task was carried out, variations during a performance task is essential to eval- the obtained findings can not be generalized for all RT uate the observed individual variability. As demonstrated paradigms. According to Stuss et al. [35], RT variability in the present study, the covariate vigilance explains the might increase in more complex tasks. Therefore, sev- variance to a certain extent. Vigilance-specific alterations eral RT tasks involving different qualities of sensory per- of electric brain activity during performance tasks ception and various levels of difficulty should be should thus be taken into account. By disregarding indi- implemented. Another drawback of the study is that the vidual vigilance-driven electrical brain signal fluctua- used version of VIGALL could not be adjusted individu- tions, relations between intra-individual variability of ally, as it used equal EEG criteria for all participants. RTs and neurological disease might be either over- or Currently, the VIGALL team is refining the operating underestimated. mode of the algorithm taking into account the indivi- Hence, the present study is methodologically impor- dual alpha peak instead of fixed frequency windows for tant by emphasising the necessity of considering vigi- VIGALL processing. Consequently, individual bound- lance in studies whilst planning, performing and aries of frequency bands could be justified, making the interpreting cognitive tasks. We determined vigilance vigilance classification of VIGALL more exact. using EEG, a well-established and non-invasive method Of course, vigilance is only one possible factor which might influence RT in real life. For example, age [36], in medicine. By this method, vigilance monitoring and alcohol [37], drugs [38], certain psychiatric [39] and classification is broadly applicable in future studies to control for intra-individual variability. somatic diseases [40] as well as distraction [41] have Furthermore, the observation that vigilance affects been shown to influence RT. Our results suggest that behavioural measures opens up perspectives to further hormones which influence the sleep-wake regulation improve the validity of neuro-imaging methods. For and therefore vigilance such as glucocorticoids, melato- instance, functional imaging studies might profit from nin and leptin as well as other hormones which influ- eliminating unexplained intra-individual variance by tak- ence these endocrine systems such as estrogens, ing different states of vigilance into account. Therefore, androgens and thyroid hormones might also play a role simultaneous usage of functional imaging methods and as influencing factor on reaction time. Therefore, we are EEG is beneficial. Technical requirements have already going to investigate the influence of these hormones on been met and PET- and fMRI-compatible EEG instru- RT and differences regarding influencing factors on RT ments are now available. Thus, the covariate vigilance in females and males in future studies. Due to fact that can be controlled for by monitoring for vigilance shifts this is a pilot study using a small group of homogenous using the VIGALL algorithm during neuro-imaging pro- healthy female subjects, we have to discuss the limita- cedures. When analysing and interpreting neuro-ima- tion that we are not able to give any data regarding ging data sets, EEG-vigilance stages could either be these mentioned additional possible influencing factors considered in the general evaluation, or only those time such as age, gender, alcohol, drugs, psychiatric and segments could be taken into account that show certain somatic diseases, distraction and hormones. Another vigilance states. Further studies are needed to assess the important issue to consider is specific methodological general impact of vigilance states on neuro-imaging approach of this study. The applied CPT was very easy and monotonous, since we intended to induce a vigi- methods. lance decline. The inter-stimulus-interval was 2000 ms. Despite these advantages of the presented study, some Furthermore, the smallest possible VIGALL analysis unit limitations also have to be mentioned. One shortcoming of our examination is the fact that, for reasons of mini- is one second. We decided to analyze the whole second mising the inter-individual variability, a rather before target presentation and not while or after target Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 7 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 8. Oken BS, Salinsky MC, Elsas SM: Vigilance, alertness, or sustained presentation, because a) the stimulus presentation attention: physiological basis and measurement. Clin Neurophysiol 2006, induces an arousal and modifies the EEG and b) the 117:1885-1901. subjects’ reaction leads to artefacts in the EEG and is 9. Jokeit H, Makeig S: Different event-related patterns of γ-band power in brain waves of fast and slow-reacting subjects. Proc Natl Acad Sci USA expected 300-500 ms after target presentation. There- 1994, 91:6339-6343. fore, our results regarding vigilance are-with good rea- 10. Delorme A, Westerfield M, Makeig S: Medial prefrontal theta bursts son-limited to the one second before target presentation. precede rapid motor responses during visual selective attention. J Neurosci 2007, 27:11949-11959. In summary, the results of the present study demon- 11. Makeig S, Jung T-P: Tonic, phasic, and transient EEG correlates of strate that RT is associated with prestimulus vigilance as auditory awareness in drowsiness. Brain Res Cogn Brain Res 1996, 4:15-25. it can be measured using EEG. By including this EEG- 12. Davis H, Davis PA, Loomis AL, Harvey EN, Hobart G: Changes in human brain potentials during the onset of sleep. Science 1937, 86:448-450. measured vigilance in the analysis, the conclusiveness of 13. Bente D: Vigilanz, dissoziative Vigilanzverschiebung und Insuffizienz des scientific data on cognitive performance or reaction tasks Vigilitätstonus. In Begleitwirkung und Mißerfolge der psychiatrischen might be improved. This may be relevant for neuropsy- Pharmakotherapie. Edited by: Kranz H, Heinrich K. Stuttgart: Thieme; 1964:13-28. chological as well as for functional neuroimaging studies. 14. Roth B: Clinical and theoretical importance of EEG rhythms corresponding to states of lowered vigilance. Electroencephalogr Clin Neurophysiol 1961, 13:395-399. Acknowledgements 15. Peiris MT, Jones RD, Davidson PR, Carroll GJ, Signal TL, Parkin PJ, van den Publication of this study was supported by the Claussen-Simon-Foundation. Berg M, Bones PJ: Identification of vigilance lapses using EEG/EOG by expert human raters. Proceedings of the Annual International Conference of Author details the IEEE Engineering in Medicine and Biology Society: 1-4 September 2005; Department of Psychiatry and Psychotherapy, University of Leipzig, Shanghai 2006, 6:5735-5737. Semmelweisstr. 10, 04103 Leipzig, Germany. Department of Psychiatry and 16. Hegerl U, Stein M, Mulert C, Mergl R, Olbrich S, Dichgans E, Rujescu D, Psychotherapy, Medical Faculty, RWTH Aachen University, Pauwelsstr. 30, Pogarell O: EEG-vigilance differences between patients with borderline 52074 Aachen, Germany. personality disorder, patients with obsessive-compulsive disorder and healthy controls. Eur Arch of Psychiatry Clin Neurosci 2005, 258:137-143. Authors’ contributions 17. Olbrich S, Sander C, Jahn I, Eplinius F, Claus S, Mergl R, Schönknecht P, The presented work was carried out in collaboration between all authors. All Hegerl U: Unstable EEG-vigilance in patients with cancer-related fatique authors were involved in drafting and revising critically the manuscript and in comparison to healthy controls. World J Biol Psychiatry . approved the final version for publication. PS and UH defined the research 18. Wittchen H-U, Zaudig M, Fydrich T: SKID-Strukturiertes Klinisches theme and planned the conception of the study. MT and CS designed the Interview für DSM-IV. Achse I und II Göttingen: Hogrefe; 1996. experiment methods and acquired data. JM and HH made substantial 19. Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro MG: The Alcohol Use contributions to the conception and design, analyzed the data, interpreted Disorders Identification Test: Guidelines for use in primary care. Geneva: the results and wrote the paper. SO and AS directed the EEG recording WHO;, 2 1989. methods and discussed analyses, interpretation and presentation. 20. Berman AH, Bergman H, Palmstierna T, Schlyter F: Evaluation of the Drug Use Disorders Identification Test (DUDIT) in criminal justice and Competing interests detoxification settings and in a Swedish population sample. Eur Addict All authors declare not to have any conflict of interest including any Res 2005, 11:22-31. financial, personal or other relationships with other people or organizations 21. Delorme A, Sejnowski T, Makeig S: Enhanced detection of artefacts in EEG that could inappropriately influence, or be perceived to influence, their work. data using higher-order statistics and independent component analysis. Neuroimage 2007, 34:1443-1449. Received: 21 April 2011 Accepted: 5 August 2011 22. Olbrich S, Jödicke J, Sander C, Himmerich H, Hegerl U: ICA-based muscle Published: 5 August 2011 artefact correction of EEG data: What is muscle and what is brain? Comment on McMenamin et al. Neuroimage 2011, 54:4-9. References 23. Schroeder MJ, Barr RE: An alpha modulation index for 1. MacDonald SW, Hultsch DF, Bunce D: Intra-individual variability in electroencephalographic studies using complex demodulation. Med Biol vigilance performance: does degrading visual stimuli mimic age-related Eng Comput 2000, 38:306-310. “neural noise"? J Clin Exp Neuropsycho 2006, 28:655-675. 24. Niedermeyer E: Alpha rhythms as physiological and abnormal 2. Reimers S, Maylor EA: Gender effects on reaction time variability and phenomena. Int J Psychophysiol 1997, 26:31-49. trial-to-trial performance: reply to Deary and Der (2005). Neuropsychol 25. De Gennaro L, Ferrara M, Curcio G, Cristiani R: Antero-posterior EEG Dev Cogn B Aging Neuropsychol Cogn 2006, 13:479-489. changes during the wakefulness-sleep transition. Clin Neurophysiol 2001, 3. Myerson J, Robertson S, Hale S: Aging and intra-individual variability in 112:1901-1911. performance: analyses of response time distributions. J Exp Anal Behav 26. Manshanden I, De Munck JC, Simon NR, Lopes da Silva FH: Source 2007, 88:319-337. localization of MEG sleep spindles and the relation to sources of alpha 4. MacDonald SW, Nyberg L, Sandblom J, Fischer H, Bäckman L: Increased band rhythms. Clin Neurophysiol 2002, 113:1937-1947. response-time variability is associated with reduced inferior parietal 27. Coatanhay A, Soufflet L, Staner L, Boeijinga P: EEG source identification: activation during episodic recognition in aging. J Cogn Neurosci 2008, frequency analysis during sleep. C R Biol 2002, 325:273-282. 20:779-786. 28. Connemann BJ, Mann K, Lange-Asschenfeldt C, Ruchsow M, 5. Hultsch DF, MacDonald SW, Hunter MA, Levy-Bencheton J, Strauss E: Intra- Schreckenberger M, Bartenstein P, Gründer G: Anterior limbic alpha-like individual variability in cognitive performance in older adults: activity: a low resolution electromagnetic tomography study with comparison of adults with mild dementia, adults with arthritis, and lorazepam challenge. Clin Neurophysiol 2005, 116:886-894. healthy adults. Neuropsychology 2000, 14:588-598. 29. Broughton R, Hasan J: Quantitative topographic electroencephalographic 6. De Frias CM, Dixon RA, Fisher N, Camicioli R: Intra-individual variability in mapping during drowsiness and sleep onset. J Clin Neurophysiol 1995, neurocognitive speed: a comparison of Parkinson’s disease and normal 12:372-386. older adults. Neuropsychologia 2007, 45:2499-2507. 30. Klimesch W: EEG alpha and theta oscillations reflect cognitive and 7. Jensen AR: The importance of intra-individual variation in reaction time. memory performance: a review and analysis. Brain Res Brain Res Rev 1999, Pers Individ Dif 1992, 13:869-881. 29:169-195. Minkwitz et al. Behavioral and Brain Functions 2011, 7:31 Page 8 of 8 http://www.behavioralandbrainfunctions.com/content/7/1/31 31. Strijkstra AM, Beersma DG, Drayer B, Halbesma N, Daan S: Subjective sleepiness correlates negatively with global alpha (8-12 Hz) and positively with central frontal theta (4-8 Hz) frequencies in the human resting awake electroencephalogram. Neurosci Lett 2003, 340:17-20. 32. Lansing RW, Schwartz E, Lindsley DB: Reaction time and EEG activation under alerted and nonalerted conditions. J Exp Psychol 1959, 58:1-7. 33. Makeig S, Inlow M: Lapses in alertness: coherence of fluctuations in performance and EEG spectrum. Electroencephalogr Clin Neurophysiol 1993, 86:23-35. 34. Jung TP, Makeig S, Stensmo M, Sejnowski TJ: Estimating alertness from the EEG power spectrum. IEEE Trans Biomed Eng 1997, 44:60-69. 35. Stuss DT, Pogue J, Buckle L, Bondar J: Characterization of stability of performance in patients with traumatic brain injury: variability and consistency on reaction time tests. Neuropsychology 1994, 8:316-324. 36. Anstey KJ, Dear K, Christensen H, Jorm AF: Biomarkers, health, lifestyle and demographic variables as correlates of reaction time performance in early, middle, and late adulthood. Q J Exp Psychol 2005, 58:5-21. 37. Howland J, Rohsenow DJ, Arnedt JT, Bliss CA, Hunt SK, Calise TV, Heeren T, Winter M, Littlefield C, Gottlieb DJ: The acute effects of caffeinated versus non-caffeinated alcoholic beverage on driving performance and attention/reaction time. Addiction 2011, 106:335-341. 38. Brockel BJ, Fowler SC: Effects of chronic haloperidol on reaction time and errors in a sustained attention task: partial reversal by anticholinergics and by amphetamine. J Pharmacol Exp Ther 1995, 275:1090-1098. 39. Kaiser S, Roth A, Rentrop M, Friederich H-C, Bender S, Weisbrod M: Intra- individual reaction time variability in schizophrenia, depression and borderline personality disorder. Brain Cogn 2008, 66:73-82. 40. Mellgren A, Norkrans G, Hagberg L, Dunlop O, Wejstal R, Gisslen M: Slowed reaction time in HIV-1-infected patients without AIDS. Acta Neurol Scand 2000, 102:169-174. 41. Travis F, Tecce JJ: Effects of distracting stimuli on CNV amplitude and reaction time. Int J Psychophysiol 1998, 31:45-50. doi:10.1186/1744-9081-7-31 Cite this article as: Minkwitz et al.: Prestimulus vigilance predicts response speed in an easy visual discrimination task. Behavioral and Brain Functions 2011 7:31. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit

Journal

Behavioral and Brain Functions – Springer Journals

Published: Aug 5, 2011

There are no references for this article.