EEG background activity described by a large computerized database
Introduction
Computer-based technology is widely applied to acquire and store electroencephalographic (EEG) data. For the electroencephalographer (EEGer), it gives the opportunity for better evaluation through user-selected montages, vertical and horizontal scaling, and filter adjustments (Blum, 1998). However, when it comes to storage and retrieval of the interpreted EEG data, the potential advantages of this new technology have not yet been exploited. Storing the EEG interpretations as free text reports leaves them inaccessible for direct statistical analysis.
Description of the background activity (BA) is essential for visual analysis of the EEG. An important part of the BA is the AR. Both AR and the rest of the BA depend on the maturity of the brain, state of alertness, and any cerebral dysfunction. In a previous paper, we have described a system for categorization of visually assessed digital EEG data in a computerized database, thus achieving accessibility of routine EEGs for statistical analysis (Aurlien et al., 1999).
Quality management has been a focused issue during the recent years. The demand for documentation of medical care addressing quality measurements and detailed information of production is increasing. Both quality control and research will profit from exploiting EEG databases. This paper demonstrates how information collected from 4651 consecutive patients over two years can be extracted in a systematic way. EEG findings can be evaluated for statistical analysis, interobserver variation can be tested, and single tests can be retrieved using specified findings in the EEGs as search criteria.
Visual analysis is still the most common way of evaluating EEGs for clinical practice. However, after the introduction of quantitative analysis and statistical evaluation of brain electrical activity (neurometrics) (Matousek et al., 1979, John and Prichep, 1993), research studies usually favour this technology. Parameters from neurometrics can usually not be directly compared with those obtained by visual analysis. Basic entities in the visual evaluation of EEGs remain undefined. EEGers often describe BA as ‘appropriate for age’. But what is appropriate for age for visually evaluated frequencies and amplitudes of the BA? The literature is rich in descriptions for AR frequency (Obrist, 1954, Otomo, 1966, Petersen and Eeg-Olofsson, 1971, Shigeta et al., 1995, Wang and Busse, 1969) but very few studies include all age groups (Hughes and Cayaffa, 1977). For the general BA apart from the AR, detailed documentation concerning visually assessed frequencies and amplitudes is sparse, and there are different conclusions whether the AR in healthy subjects declines in older age or not.
The main purpose of this paper is to show how our newly developed software connected to an EEG database gives easy access to the findings obtained from the visual analysis of a large number of prospective EEGs collected in a busy clinical practice. The focus will be on the visually assessed BA including the AR.
Section snippets
EEG recordings
All 5976 EEGs recorded at Haukeland University Hospital from January 1, 2000 to March 3, 2002 were visually evaluated and described. Long-term registrations and EEGs during WADA tests and Tilt tests were not included in this study. The EEGs were described by one of 5 EEGers. EEGer 1 described 1044 EEGs, EEGer 2 described 1947 EEGs, EEGer 3 described 1624 EEGs, EEGer 4 described 1309 EEGs, and EEGer 5 described 52 EEGs. Recordings from EEGer 5 were excluded from interobserver variability testing
EEG events
EEG events were described in 3005 EEGs. Of the total 4651 EEGs included, 2867 (62%) were evaluated as normal and 1784 (38%) as pathological. 637 EEGs (14%) contained epileptiform pathology.
AR frequency
AR frequencies showed a wide variation in all age groups. The fitted polynomial model was of 7th order. Estimated mean AR frequency increased gradually until age 20 years to a value of 10 Hz (Fig. 2A). The frequency remained stable until age 45 years and then declined. Higher AR frequencies were recorded for
Discussion
This paper illustrates how our recently developed software for interactive systematization and categorization of EEG interpretation makes it possible to access a large clinical material of EEGs. The data were collected prospectively for 2 years, and the system proved feasible in a daily routine without use of extra resources.
Quality management has been a focused issue during recent years. The interobserver variability is an important factor for quality control measurements (Walczak et al., 1992
References (27)
- et al.
A new way of building a database of EEG findings
Clin Neurophysiol
(1999) Computer-based electroencephalography: technical basics, basis for new applications, and potential pitfalls
Electroenceph clin Neurophysiol
(1998)- et al.
EEG and event-related potentials in normal aging
Prog Neurobiol
(1993) - et al.
Life-span changes in EEG spectral amplitude, amplitude variability and mean frequency
Clin Neurophysiol
(1999) - et al.
The EEG in patients at different ages without organic cerebral disease
Electroenceph clin Neurophysiol
(1977) - et al.
Serial quantitative electroencephalography
Electroenceph clin Neurophysiol
(1979) Alpha rhythms as physiological and abnormal phenomena
Int J Psychophysiol
(1997)The electroencephalogram of normal aged adults
Electroenceph clin Neurophysiol
(1954)Electroencephalography in old age: dominant alpha pattern
Electroenceph clin Neurophysiol
(1966)EEG and ERP assessment of normal aging
Electroenceph clin Neurophysiol
(1997)
EEG in successful aging; a 5 year follow-up study from the 8h to ninth decade of life
Electroenceph clin Neurophysiol
Table X2.5. EEG waveform/activity identifiers. ASTM Designation: E1467-94. Standard specifications for transferring digital neurophysiological data between independent computer systems
An analysis of transformations
J R Statist Soc B
Cited by (123)
Effect of road markings and road signs quality on driving behaviour, driver's gaze patterns and driver's cognitive load at night-time
2023, Transportation Research Part F: Traffic Psychology and BehaviourOptimization enabled deep residual neural network for motor imagery EEG signal classification
2023, Biomedical Signal Processing and ControlCitation Excerpt :Complicated processing of the signal is modelled with codes for converting it to false actuator signal for managing exterior tools based on the flow of design as defined in [13]. The recorded EEG signal is buried by several artifacts based on physiological and non-physiological sources [14]. Thus, these artifacts must be eliminated carefully without destroying the beneficial data with filtering techniques [15].
Fully automatic peak frequency estimation of the posterior dominant rhythm in a large retrospective hospital EEG cohort
2021, Clinical Neurophysiology Practice