Elsevier

Clinical Neurophysiology

Volume 110, Issue 9, 1 September 1999, Pages 1499-1509
Clinical Neurophysiology

Vigilance transitions in reaction time test: a method of describing the state of alertness more objectively

https://doi.org/10.1016/S1388-2457(99)00115-7Get rights and content

Abstract

Electrophysiological recordings are considered a reliable method of assessing a person's alertness. The aim of this study was to show, firstly, that changes in alertness during a Reaction Time Test (RTT) can be determined with certain adaptive scoring stages but not with R&K scoring and secondly, that the different adaptive stages can explain findings in reaction time. In 17 male patients (50.8±9.7 years, Body-Mass Index (BMI) 31.9±5.1 kg/m2) diagnosed with Obstructive Sleep Apnea Syndrome (OSAS) (Respiratory Disturbance Index (RDI) 53.3±24.1 /h sleep) a 90 min daytime vigilance test was performed twice, after the diagnostic polysomnographic investigation and after two nights spent with nCPAP. After a computerised adaptive segmentation analysis, a visual rule-defined classification system categorised alertness into one of 12 adaptive scoring stages. 6 of the 12 stages are described by the alertness conditions comparable to WAKE and NREM1. 4 stages are nearly classified as NREM2-4, Rapid Eye Movement (REM) and Movement Time (MT), and one stage reflects the increase of alertness from drowsiness. The typical stage of an alert subject increased significantly from a median of 65.9% before therapy to 80.8% in the second investigation. The percentages of clearly drowsy stages decreased significantly. In contrast, there were no significant changes in the percentages of sleep stages according to R&K criteria for both investigations. According to R&K criteria 178 of 398 failed reactions (Reaction time>10 s) occurred in stage WAKE. According to adaptive scoring, only 12 failed reactions appeared in the alert stage. During the other failed reactions the electrophysiological recordings showed decreases in alertness. Neither the visual assessment nor the descriptive statistical results of R&K scoring were helpful to interpret the patient's alertness condition. In contrast, the patients’ increases in alertness with nCPAP could be described by the adaptive scoring stages. This method could be a very useful procedure, when an expert opinion is necessary. It also has an actual context to the discussion of the effectiveness of CPAP in the treatment of OSAS.

Introduction

Sleep medicine is asked to offer objective methods to measure daytime alertness, tiredness and sleepiness. However, to describe patients’ ability in falling asleep during the daytime, especially in the diagnosis of narcolepsy, the Multiple Sleep Latency Test (MSLT) is well described and standardised (Carskadon et al., 1986). In various sleep–wake disturbances the main question is not whether a patient can relax and fall asleep after previous instruction, as in the MSLT. Therefore it is not an adequate procedure to describe the patient's attention over a longer period of time and the ability to react adequately to an internal or an external stimuli.

The condition of alertness may need to be quantified because subjective reports are imprecise compared with observer reports or tests of sleepiness (Walsleben, 1992, Kribbs et al., 1993). The subjective estimation of alertness or drowsiness is of limited value as it is unreliable and it is difficult to use comparisons of different subjects (Hoddes et al., 1972, Dement et al., 1978). Most psychometric investigations have two disadvantages. Firstly, most of these tests have an activating character for the patients. The second disadvantage in most psychometric investigations are the missing electrophysiological recordings (EEG, EOG, EMG). The electrophysiology is a reliable method of detecting whether a person is awake, tired or asleep. The Rechtschaffen and Kales (1968) (R&K) criteria are the best internationally recognised standard for sleep analysis. Initially, they were used to reference scorings in different laboratories. Today it is used for almost all sleep studies. The standard epoch length of 30 s is too long for the description of short micro-sleep episodes or fast fluctuations in alertness. The electrophysiological changes related to drowsiness vary rapidly (Roth, 1961, Oswald, 1962, Santamaria and Chiappa, 1987). Shorter epoch durations have been used in previous studies dealing with the relationship between electrophysiological activity, alertness and reactivity (Morell, 1966, Matusek and Petersén, 1983, Belyavin and Wright, 1987, Torsvall and Åkerstedt, 1987). The R&K criteria have no subdivisions for description of drowsiness or the sleep–wake transitions. Subdivisions of wakefulness and drowsiness are important because even minor changes in the electrophysiological state have been found to correlate with differences in reactivity and performance (Morell, 1966, O'Hanlon and Kelly, 1977, O'Hanlon and Beatty, 1977, Torsvall and Åkerstedt, 1987, Broughton et al., 1988).

In our investigation an adaptive scoring system was used which had no distinct epoch lengths but the epoch boundaries had been determined automatically by properties of the signals (Krajca et al., 1991, Värri et al., 1992). As long as the electrophysiological activity remained constant the same epoch continued and as soon as a change took place the epoch changed. This adaptive scoring classification included a more detailed division of wakefulness and drowsiness (based on Hasan et al., 1993). Since the drops in a patient's state of alertness can have a short duration and fast transitions, this method is expected to correspond more closely to the electrophysiological state of humans. In our investigation a simple 90 min Four Choice Reaction Time Test (FCRTT) with electrophysiological recordings was performed to evaluate patients’ daytime alertness and reactivity. This test has been sensitive to treatment of Obstructive Sleep Apnea Syndrome (OSAS) by a change to a shorter reaction time in several previous studies (Cassel et al., 1996, Conradt et al., 1998).

The aim of this study was to show, firstly, that changes in alertness during the FCRTT can be determined with certain adaptive stages but not with R&K scoring and secondly, that the adaptive segmentation in distinct stages can explain findings in reaction time. In patients with OSAS-impaired alertness and Excessive Daytime Sleepiness (EDS) result in daily disturbances (Guilleminault et al., 1993). OSAS is one of the most common medical causes of moderate and severe sleepiness (American Thoracic Society Board of Directors Sleep Apnea, 1994). Nasal CPAP is the most common and effective treatment of OSAS (He et al., 1988). In these patients, the state of daytime alertness improves rapidly with nCPAP therapy, as compared with pre-nCPAP. Therefore, these daytime variances are a good model demonstrating the changes in alertness in the electrophysiological recordings.

Section snippets

Patients

17 male patients were included based on the following criteria: Caucasian, diagnosed OSAS with an RDI>20/h and the presence of subjective impaired daytime performance with an indication for nCPAP therapy and the willingness to participate. The mean age of the 17 subjects was 50.8±9.7 years, the Body-Mass Index (BMI) 31.9±5.1 kg/m2 and the mean Respiratory Disturbance Index (RDI) amounted to 53.3±24.1 events per hour of sleep. A complete physical examination was performed and a full medical

Results

OSAS was treated effectively in each of the 17 patients after the second night spent with an effective nCPAP pressure. The mean RDI decreased from 53.3±24.1 to 1.6±1.8 with effective nCPAP therapy. The mean total sleep time did not change significantly from 357.7±50.7 min before therapy to 387.3±45.0 min on the second night spent with nCPAP. The subjective estimation of daytime alertness was markedly increased in 16 of the 17 patients. The ESS score decreased from a median of 12 (10–18) before

Discussion

It is of current interest to develop a suitable method which describes performance during daytime testing as precisely as possible. Changes in alertness are important tools used to describe therapeutic effects in scientific investigations. With a more practical background, a frequent question for the expert is the restoration of working ability in patients with therapy.

The criteria for the different adaptive scoring stages used in the present work are based on established electrophysiological

Acknowledgements

We thank Monique Aarts from Sydney for her linguistic competence.

References (33)

  • A. Värri et al.

    Non-linear eye movement detection method for drowsiness studies

    Int J Bio-Med Comput

    (1996)
  • S. Volk et al.

    The structure of wakefulness and its relationship to daytime sleep in narcoleptic patients

    Electroenceph clin Neurophysiol

    (1984)
  • J.A. Walsleben

    The measurement of daytime wakefulness (editorial)

    Chest

    (1992)
  • Sleepiness and Driving Risk

    Am J Crit Care Med

    (1994)
  • M.A. Carskadon et al.

    Guidelines for the multiple sleep latency test (MSLT): a standard measure of sleepiness

    Sleep

    (1986)
  • W. Cassel et al.

    Risk of traffic accidents in patients with sleep disordered breathing: reduction with nCPAP

    Eur Respir J

    (1996)
  • Cited by (16)

    • Technology to Detect Driver Sleepiness

      2019, Sleep Medicine Clinics
      Citation Excerpt :

      The test was chosen to be so long that patients used to performing tasks when being sleepy and used to fighting against sleepiness still might become more sleepy and may fall asleep during a boring task.20,21 For this kind of evaluation, a more detailed classification for the EEG with vigilance stages between bright awake and falling asleep had been defined.19 This classification is based mainly on the disappearance of beta waves, appearance and disappearance of alpha waves, and the first appearance of theta waves.

    • Application of adaptive neuro-fuzzy inference system for vigilance level estimation by using wavelet-entropy feature extraction

      2009, Expert Systems with Applications
      Citation Excerpt :

      Therefore, for transient signals such as EEG, the wavelet analysis is superior to the Fourier transform. DWT analysis of EEG recordings was proven to be a powerful tool for detecting transitions from an alert to a drowsy state (Conradt et al., 1999; Jung et al., 1997). The DWT analysis used the fact that such an EEG recording comprises a characteristic rhythm that would disappear when a subject becomes drowsy.

    • The ANN-based computing of drowsy level

      2009, Expert Systems with Applications
    View all citing articles on Scopus
    View full text