Using normative modeling and machine learning for detecting mild traumatic brain injury from magnetoencephalography data.

New biomarkers are urgently needed for many brain disorders; for example, the diagnosis of mild traumatic brain injury (mTBI) is challenging as the clinical symptoms are diverse and nonspecific. EEG and MEG studies have demonstrated several population-level indicators of mTBI that could serve as objective markers of brain injury. However, deriving clinically useful biomarkers for mTBI and other brain disorders from EEG/MEG signals is hampered by the large inter-individual variability even across healthy people. Here, we used a multivariate machine-learning approach to detect mTBI from resting-state MEG measurements. To address the heterogeneity of the condition, we employed a normative modeling approach and modeled MEG signal features of individual mTBI patients as deviations with respect to the normal variation. To this end, a normative dataset comprising 621 healthy participants was used to determine the variation in power spectra across the cortex. In addition, we constructed normative datasets based on age-matched subsets of the full normative data. To discriminate patients from healthy control subjects, we trained support-vector-machine classifiers on the quantitative deviation maps for 25 mTBI patients and 20 controls not included in the normative dataset. The best performing classifier made use of the full normative data across the entire age and frequency ranges. This classifier was able to distinguish patients from controls with an accuracy of 79%. Inspection of the trained model revealed that low-frequency activity in the theta frequency band (4-8 Hz) is a significant indicator of mTBI, consistent with earlier studies. The results demonstrate the feasibility of using normative modeling of MEG data combined with machine learning to advance diagnosis of mTBI and identify patients that would benefit from treatment and rehabilitation. The current approach could be applied to a wide range of brain disorders, thus providing a basis for deriving MEG/EEG-based biomarkers.

Sensor-level MEG combined with machine learning yields robust classification of mild traumatic brain injury patients.

OBJECTIVE: Diagnosis of mild traumatic brain injury (mTBI) is challenging despite its high incidence, due to the unspecificity and variety of symptoms and the frequent lack of structural imaging findings. There is a need for reliable and simple-to-use diagnostic tools that would be feasible across sites and patient populations. METHODS: We evaluated linear machine learning (ML) methods' ability to separate mTBI patients from healthy controls, based on their sensor-level magnetoencephalographic (MEG) power spectra in the subacute phase (<2 months) after a head trauma. We recorded resting-state MEG data from 25 patients and 25 age-sex matched controls and utilized a previously collected data set of 20 patients and 20 controls from a different site. The data sets were analyzed separately with three ML methods. RESULTS: The median classification accuracies varied between 80 and 95%, without significant differences between the applied ML methods or data sets. The classification accuracies were significantly higher with ML than with traditional sensor-level MEG analysis based on detecting pathological low-frequency activity. CONCLUSIONS: Easily applicable linear ML methods provide reliable and replicable classification of mTBI patients using sensor-level MEG data. SIGNIFICANCE: Power spectral estimates combined with ML can classify mTBI patients with high accuracy and have high promise for clinical use.

Mild Traumatic Brain Injury Affects Cognitive Processing and Modifies Oscillatory Brain Activity during Attentional Tasks.

Despite the high prevalence of mild traumatic brain injury (mTBI), current diagnostic tools to objectively assess cognitive complaints after mTBI continue to be inadequate. Our aim was to identify neuronal correlates for cognitive difficulties in mTBI patients by evaluating the possible alterations in oscillatory brain activity during a behavioral task known to be sensitive to cognitive impairment after mTBI. We compared oscillatory brain activity during rest and cognitive tasks (Paced Auditory Serial Addition Test [PASAT] and a vigilance test [VT]) with magnetoencephalography between 25 mTBI patients and 20 healthy controls. Whereas VT induced no significant differences compared with resting state in either group, patients exhibited stronger attenuation of 8- to 14-Hz oscillatory activity during PASAT than healthy controls in the left parietotemporal cortex (p ≤ 0.05). Further, significant task-related modulation in the left superior frontal gyrus and right prefrontal cortex was detected only in patients. The ∼10-Hz (alpha) peak frequency declined in frontal, temporal, and parietal regions during PASAT compared with rest (p < 0.016) in patients, whereas in controls it remained the same or showed a tendency to increase. In patients, the ∼10-Hz peak amplitude was negatively correlated with behavioral performance in the Trail Making Test. The observed alterations in the cortical oscillatory activity during cognitive load may provide measurable neurophysiological correlates of cognitive difficulties in mTBI patients, even at the individual level.

Theta-Band Oscillations as an Indicator of Mild Traumatic Brain Injury.

Mild traumatic brain injury (mTBI) patients continue to pose a diagnostic challenge due to their diverse symptoms without trauma-specific changes in structural imaging. We addressed here the possible early changes in spontaneous oscillatory brain activity after mTBI, and their feasibility as an indicator of injury in clinical evaluation. We recorded resting-state magnetoencephalography (MEG) data in both eyes-open and eyes-closed conditions from 26 patients (11 females and 15 males, aged 20-59) with mTBI 6 days-6 months after the injury, and compared their spontaneous oscillatory activity to corresponding data from 139 healthy controls. Twelve of the patients underwent a follow-up measurement at 6 months. Ten of all patients were without structural lesions in MRI. At single-subject level, aberrant 4-7 Hz (theta) band activity exceeding the + 2 SD limit of the healthy subjects was visible in 7 out of 26 patients; three out of the seven patients with abnormal theta activity were without any detectable lesions in MRI. Of the patients that participated in the follow-up measurements, five showed abnormal theta activity in the first recording, but only two in the second measurement. Our results suggest that aberrant theta-band oscillatory activity can provide an early objective sign of brain dysfunction after mTBI. In 3/7 patients, the slow-wave activity was transient and visible only in the first recording, urging prompt timing for the measurements in clinical settings.

Slow-Wave Oscillations in Awake Healthy Subjects: Methodological and Physiological Considerations.

PURPOSE: Detection of pathologic slow-wave oscillations (0.5-7 Hz) in awake subjects has gained increasing interest in clinical diagnostics. Their significance, however, is hampered by the occasional presence of slow waves in healthy subjects, as well as the abundance of artefactual signals at low measurement frequencies. The aim of this study was to assess the occurrence of slow-wave oscillations in healthy subjects and to sharpen the management of possible measurement artifacts, in order to create a normative database for neurological patients. METHODS: The authors analyzed magnetoencephalography recordings of spontaneous brain oscillations in 139 awake healthy adults. Sources of artifacts were first identified and suppressed by temporal extension of signal space separation method, and the remaining artifact components were projected out using signal space projection. Individual amplitude spectra were compared with the channel-level average spectra over all subjects. RESULTS: Slow-wave oscillations deviating ±2 standard deviations from the average spectrum were detected in 12 subjects (∼9%). In 10 subjects, the oscillations were considered as normal physiological phenomena. Only two subjects showed activity that could have been interpreted as pathological: one subject with widespread parietal bilateral polyrhythmic slow-wave activity and one with focal rolandic 2.7-Hz slow-wave activity. CONCLUSIONS: The prevalence of slow-wave oscillations in a healthy adult population is low. Knowledge about their occurrence, however, is essential for interpreting their significance in brain diseases. Artifacts and benign oscillatory variants at slow frequencies have to be recognized.

[Aseptic meningitis]. / Aseptinen meningiitti.

Aseptic meningitis is a benign condition often triggered by a virus or an immunological process. For example herpes virus, borrelia, tuberculosis, a fungus or an autoimmune disease may underlie meningitides presenting prolonged or recurrent symptoms. It is essential to identify the meningitis patients among the diverse group of headache patients and carry out focused investigations and treatment, and in mild cases to avoid complications caused by the investigations. Analgesic and antiemetic medication are usually sufficient for symptomatic treatment. Etiological treatment is available for some patients.