Jussi Korpela

Mental workload classification using heart rate metrics

The ability of different short-term heart rate variability metrics to classify the level of mental workload (MWL) in 140 s segments was studied. Electrocardiographic data and event related potentials (ERPs), calculated from electroencephalographic data, were collected from 13 healthy subjects during the performance of a computerised cognitive multitask test with different task load levels. The amplitude of the P300 component of the ERPs was used as an objective measure of MWL. Receiver operating characteristics analysis (ROC) showed that the time domain metric of average interbeat interval length was the best-performing metric in terms of classification ability.

Improving the saccade peak velocity measurement for detecting fatigue.

The aim of the study was to compare saccadic peak velocity (SPV) values measured with video based Fitness Impairment Tester (FIT) and electro-oculography (EOG) during prolonged wakefulness. We tested different numbers of saccades and two saccade paradigms to improve the EOG measurements for detecting fatigue. The SPVs were measured from 11 fast patrol boat navigators with FIT and EOG every sixth hour until 54 h. Subjective sleepiness was assessed with the Karolinska Sleepiness Scale. EOG was measured using an overlap and a gap paradigm and the data was divided into sequential five 20-saccade blocks and cumulative blocks of 20, 40, 60, 80, and 100 saccades. Compared to the gap paradigm, the overlap paradigm produced a higher number of analyzable saccades for a given measurement time. The shorter measurements (20-40 saccades) appeared to be more sensitive for fatigue, whereas the longer measurements (60-100 saccades) were more sensitive to time spent on the task. Thus, the optimal number of saccades varies also depending on the research question. The EOG method was more sensitive to fatigue than FIT. The FIT values measured after 30 and 36 h of wakefulness did not differ significantly from the baseline values, while subjective sleepiness and the EOG values showed that the participants were significantly less alert at these time points. The EOG measurements can be improved for detecting fatigue by using the overlap saccade paradigm. The SPV values measured with the EOG method appear to be somewhat more sensitive in detecting fatigue than the FIT method.

Alterations in attention capture to auditory emotional stimuli in job burnout: An event-related potential study

Job burnout is a significant cause of work absenteeism. Evidence from behavioral studies and patient reports suggests that job burnout is associated with impairments of attention and decreased working capacity, and it has overlapping elements with depression, anxiety and sleep disturbances. Here, we examined the electrophysiological correlates of automatic sound change detection and involuntary attention allocation in job burnout using scalp recordings of event-related potentials (ERP). Volunteers with job burnout symptoms but without severe depression and anxiety disorders and their non-burnout controls were presented with natural speech sound stimuli (standard and nine deviants), as well as three rarely occurring speech sounds with strong emotional prosody. All stimuli elicited mismatch negativity (MMN) responses that were comparable in both groups. The groups differed with respect to the P3a, an ERP component reflecting involuntary shift of attention: job burnout group showed a shorter P3a latency in response to the emotionally negative stimulus, and a longer latency in response to the positive stimulus. Results indicate that in job burnout, automatic speech sound discrimination is intact, but there is an attention capture tendency that is faster for negative, and slower to positive information compared to that of controls.

Job burnout is associated with dysfunctions in brain mechanisms of voluntary and involuntary attention

Individuals with job burnout symptoms often report having cognitive difficulties, but related electrophysiological studies are scarce. We assessed the impact of burnout on performing a visual task with varying memory loads, and on involuntary attention switch to distractor sounds using scalp recordings of event-related potentials (ERPs). Task performance was comparable between burnout and control groups. The distractor sounds elicited a P3a response, which was reduced in the burnout group. This suggests burnout-related deficits in processing novel and potentially important events during task performance. In the burnout group, we also observed a decrease in working-memory related P3b responses over posterior scalp and increase over frontal areas. These results suggest that burnout is associated with deficits in cognitive control needed to monitor and update information in working memory. Successful task performance in burnout might require additional recruitment of anterior regions to compensate the decrement in posterior activity.

Fast determination of MMN and P3a responses to linguistically and emotionally relevant changes in pseudoword stimuli.

We developed a new multi-feature mismatch negativity (MMN) paradigm with two improvements: Firstly, the standard tone, a pseudoword /ta-ta/ was presented with equal probability to the nine linguistically relevant deviants, reducing the recording time by 45%. Secondly, three rare, emotionally valenced stimuli: happy, angry, and sad utterances of the standard pseudoword were included in the sequence. MMN signals reflecting the perceptual properties of the sounds were observed for all stimuli. In addition, P3a signals were observed for the rare emotionally uttered pseudowords. This 28-min paradigm allows a multi-dimensional evaluation of central speech-sound representations (MMN), and attention allocation (P3a) to emotional information content of speech. We recommend this paradigm for studies on subject groups with impairments in language or emotional information processing, such as autism spectrum disorders, attention disorders, and alexithymia.

Shifting of attentional set is inadequate in severe burnout: Evidence from an event-related potential study

Individuals with prolonged occupational stress often report difficulties in concentration. Work tasks often require the ability to switch back and forth between different contexts. Here, we studied the association between job burnout and task switching by recording event-related potentials (ERPs) time-locked to stimulus onset during a task with simultaneous cue-target presentation and unpredictable switches in the task. Participants were currently working people with severe, mild, or no burnout symptoms. In all groups, task performance was substantially slower immediately after task switch than during task repetition. However, the error rates were higher in the severe burnout group than in the mild burnout and control groups. Electrophysiological data revealed an increased parietal P3 response for the switch trials relative to repetition trials. Notably, the response was smaller in amplitude in the severe burnout group than in the other groups. The results suggest that severe burnout is associated with inadequate processing when rapid shifting of attention between tasks is required resulting in less accurate performance.

Computational Testing for Automated Preprocessing 2: Practical Demonstration of a System for Scientific Data-Processing Workflow Management for High-Volume EEG

Existing tools for the preprocessing of EEG data provide a large choice of methods to suitably prepare and analyse a given dataset. Yet it remains a challenge for the average user to integrate methods for batch processing of the increasingly large datasets of modern research, and compare methods to choose an optimal approach across the many possible parameter configurations. Additionally, many tools still require a high degree of manual decision making for, e.g., the classification of artifacts in channels, epochs or segments. This introduces extra subjectivity, is slow, and is not reproducible. Batching and well-designed automation can help to regularize EEG preprocessing, and thus reduce human effort, subjectivity, and consequent error. The Computational Testing for Automated Preprocessing (CTAP) toolbox facilitates: (i) batch processing that is easy for experts and novices alike; (ii) testing and comparison of preprocessing methods. Here we demonstrate the application of CTAP to high-resolution EEG data in three modes of use. First, a linear processing pipeline with mostly default parameters illustrates ease-of-use for naive users. Second, a branching pipeline illustrates CTAPs support for comparison of competing methods. Third, a pipeline with built-in parameter-sweeping illustrates CTAPs capability to support data-driven method parameterization. CTAP extends the existing functions and data structure from the well-known EEGLAB toolbox, based on Matlab, and produces extensive quality control outputs. CTAP is available under MIT open-source licence from https://github.com/bwrc/ctap.

Explaining Interval Sequences by Randomization

Sequences of events are an ubiquitous form of data. In this paper, we show that it is feasible to present an event sequence as an interval sequence. We show how sequences can be efficiently randomized, how to choose a correct null model and how to use randomizations to derive confidence intervals. Using these techniques, we gain knowledge of the temporal structure of the sequence. Time and Fourier space representations, autocorrelations and arbitrary features can be used as constraints in investigating the data. The methods presented are applied to two real-life datasets; a medical heart interbeat interval dataset and a word dataset from a book. We find that the interval sequence representation and randomization methods provide a powerful way to explore interval sequences and explain their structure.

Visual ERP P3 amplitude and latency in standalone and embedded visual processing task

Event-related potentials (ERPs) of a visual processing task are compared with and without a simultaneous external working memory load. Ten adults participated the same measurement session on three separate days. Results for visual ERP P3 amplitude and reaction time (RT) are presented for both task conditions. Both the reaction time of the visual task and the respective P3 latency increased during high memory load. It was also found that P3 amplitude and reaction time correlated only under the high memory load condition. The results indicate that visual P3 to a simple processing task is affected by external working memory load.

Using regression makes extraction of shared variation in multiple datasets easy

In many data analysis tasks it is important to understand the relationships between different datasets. Several methods exist for this task but many of them are limited to two datasets and linear relationships. In this paper, we propose a new efficient algorithm, termed cocoreg, for the extraction of variation common to all datasets in a given collection of arbitrary size. cocoreg extends redundancy analysis to more than two datasets, utilizing chains of regression functions to extract the shared variation in the original data space. The algorithm can be used with any linear or non-linear regression function, which makes it robust, straightforward, fast, and easy to implement and use. We empirically demonstrate the efficacy of shared variation extraction using the cocoreg algorithm on five artificial and three real datasets.