Decoding EEG rhythms offline and online during motor imagery for standing and sitting based on a brain-computer interface.

Motor imagery (MI)-based brain-computer interface (BCI) systems have shown promising advances for lower limb motor rehabilitation. The purpose of this study was to develop an MI-based BCI for the actions of standing and sitting. Thirty-two healthy subjects participated in the study using 17 active EEG electrodes. We used a combination of the filter bank common spatial pattern (FBCSP) method and the regularized linear discriminant analysis (RLDA) technique for decoding EEG rhythms offline and online during motor imagery for standing and sitting. The offline analysis indicated the classification of motor imagery and idle state provided a mean accuracy of 88.51 ± 1.43% and 85.29 ± 1.83% for the sit-to-stand and stand-to-sit transitions, respectively. The mean accuracies of the sit-to-stand and stand-to-sit online experiments were 94.69 ± 1.29% and 96.56 ± 0.83%, respectively. From these results, we believe that the MI-based BCI may be useful to future brain-controlled standing systems.

Machine learning in the loop for tuberculosis diagnosis support.

The use of machine learning (ML) for diagnosis support has advanced in the field of health. In the present paper, the results of studying ML techniques in a tuberculosis diagnosis loop in a scenario of limited resources are presented. Data are analyzed using a tuberculosis (TB) therapy program at a health institution in a main city of a developing country using five ML models. Logistic regression, classification trees, random forest, support vector machines, and artificial neural networks are trained under physician supervision following physicians' typical daily work. The models are trained on seven main variables collected when patients arrive at the facility. Additionally, the variables applied to train the models are analyzed, and the models' advantages and limitations are discussed in the context of the automated ML techniques. The results show that artificial neural networks obtain the best results in terms of accuracy, sensitivity, and area under the receiver operating curve. These results represent an improvement over smear microscopy, which is commonly used techniques to detect TB for special cases. Findings demonstrate that ML in the TB diagnosis loop can be reinforced with available data to serve as an alternative diagnosis tool based on data processing in places where the health infrastructure is limited.

Preliminary Text Analysis from Medical Records for TB Diagnosis Support.

Tuberculosis is an infectious disease that is spread through the air from one person to another and is one of the top ten causes of death in the world according to the World Health Organization. From biomedical engineering, decision support systems based on artificial intelligence have shown advantages for healthcare personnel in tasks such as diagnosis and screening. A specific area of the artificial intelligence is the natural language processing, however, most of these approaches are based on available data. This paper shows the construction of a dataset based on medical records of subjects suspected of tuberculosis. In addition, an initial exploration of the contents of the constructed dataset and how this approach can be followed by a natural language processing to support tuberculosis diagnosis in data demanding scenarios are presented.Clinical Relevance- In some developing countries as Colombia, it is difficult to develop systems based on artificial intelligence due to the availability of data. This proposal holds a strategy to build a dataset to train machine learning models, and to obtain support diagnosis tools, employing natural language from the medical scenario from text written by health professionals in the medical record. In this way, trained models based on this information available can be employed in places where medical infrastructure is precarious.

Sleep apnea: Tracking effects of a first session of CPAP therapy by means of Granger causality.

Connectivity between physiological networks is an issue of particular importance for understanding the complex interaction brain-heart. In the present study, this interaction was analyzed in polysomnography recordings of 28 patients diagnosed with obstructive sleep apnea (OSA) and compared with a group of 10 control subjects. Electroencephalography and electrocardiography signals from these polysomnography time series were characterized employing Granger causality computation to measure the directed connectivity among five brain waves and three spectral subbands of heart rate variability. Polysomnography data from OSA patients were recorded before and during a first session of continuous positive air pressure (CPAP) therapy in a split-night study. Results showed that CPAP therapy allowed the recovery of inner brain connectivities, mainly in subsystems involving the theta wave. In addition, differences between control and OSA patients were established in connections that involve lower frequency ranges of heart rate variability. This information can be potentially useful in the initial diagnosis of OSA, and determine the role of cardiac activity in sleep dynamics based on the use of three subbands of heart rate variability.

Brain Activity Correlates With Cognitive Performance Deterioration During Sleep Deprivation.

We studied the correlation between oscillatory brain activity and performance in healthy subjects performing the error awareness task (EAT) every 2 h, for 24 h. In the EAT, subjects were shown on a screen the names of colors and were asked to press a key if the name of the color and the color it was shown in matched, and the screen was not a duplicate of the one before ("Go" trials). In the event of a duplicate screen ("Repeat No-Go" trial) or a color mismatch ("Stroop No-Go" trial), the subjects were asked to withhold from pressing the key. We assessed subjects' (N = 10) response inhibition by measuring accuracy of the "Stroop No-Go" (SNGacc) and "Repeat No-Go" trials (RNGacc). We assessed their reactivity by measuring reaction time in the "Go" trials (GRT). Simultaneously, nine electroencephalographic (EEG) channels were recorded (Fp2, F7, F8, O1, Oz, Pz, O2, T7, and T8). The correlation between reactivity and response inhibition measures to brain activity was tested using quantitative measures of brain activity based on the relative power of gamma, beta, alpha, theta, and delta waves. In general, response inhibition and reactivity reached a steady level between 6 and 16 h of sleep deprivation, which was followed by sustained impairment after 18 h. Channels F7 and Fp2 had the highest correlation to the indices of performance. Measures of response inhibition (RNGacc and SNGacc) were correlated to the alpha and theta waves' power for most of the channels, especially in the F7 channel (r = 0.82 and 0.84, respectively). The reactivity (GRT) exhibited the highest correlation to the power of gamma waves in channel Fp2 (0.76). We conclude that quantitative measures of EEG provide information that can help us to better understand changes in subjects' performance and could be used as an indicator to prevent the adverse consequences of sleep deprivation.

Electrodermal Activity Is Sensitive to Cognitive Stress under Water.

When divers are at depth in water, the high pressure and low temperature alone can cause severe stress, challenging the human physiological control systems. The addition of cognitive stress, for example during a military mission, exacerbates the challenge. In these conditions, humans are more susceptible to autonomic imbalance. Reliable tools for the assessment of the autonomic nervous system (ANS) could be used as indicators of the relative degree of stress a diver is experiencing, which could reveal heightened risk during a mission. Electrodermal activity (EDA), a measure of the changes in conductance at the skin surface due to sweat production, is considered a promising alternative for the non-invasive assessment of sympathetic control of the ANS. EDA is sensitive to stress of many kinds. Therefore, as a first step, we tested the sensitivity of EDA, in the time and frequency domains, specifically to cognitive stress during water immersion of the subject (albeit with their measurement finger dry for safety). The data from 14 volunteer subjects were used from the experiment. After a 4-min adjustment and baseline period after being immersed in water, subjects underwent the Stroop task, which is known to induce cognitive stress. The time-domain indices of EDA, skin conductance level (SCL) and non-specific skin conductance responses (NS.SCRs), did not change during cognitive stress, compared to baseline measurements. Frequency-domain indices of EDA, EDASymp (based on power spectral analysis) and TVSymp (based on time-frequency analysis), did significantly change during cognitive stress. This leads to the conclusion that EDA, assessed by spectral analysis, is sensitive to cognitive stress in water-immersed subjects, and can potentially be used to detect cognitive stress in divers.

Highly sensitive index of sympathetic activity based on time-frequency spectral analysis of electrodermal activity.

Time-domain indices of electrodermal activity (EDA) have been used as a marker of sympathetic tone. However, they often show high variation between subjects and low consistency, which has precluded their general use as a marker of sympathetic tone. To examine whether power spectral density analysis of EDA can provide more consistent results, we recently performed a variety of sympathetic tone-evoking experiments (43). We found significant increase in the spectral power in the frequency range of 0.045 to 0.25 Hz when sympathetic tone-evoking stimuli were induced. The sympathetic tone assessed by the power spectral density of EDA was found to have lower variation and more sensitivity for certain, but not all, stimuli compared with the time-domain analysis of EDA. We surmise that this lack of sensitivity in certain sympathetic tone-inducing conditions with time-invariant spectral analysis of EDA may lie in its inability to characterize time-varying dynamics of the sympathetic tone. To overcome the disadvantages of time-domain and time-invariant power spectral indices of EDA, we developed a highly sensitive index of sympathetic tone, based on time-frequency analysis of EDA signals. Its efficacy was tested using experiments designed to elicit sympathetic dynamics. Twelve subjects underwent four tests known to elicit sympathetic tone arousal: cold pressor, tilt table, stand test, and the Stroop task. We hypothesize that a more sensitive measure of sympathetic control can be developed using time-varying spectral analysis. Variable frequency complex demodulation, a recently developed technique for time-frequency analysis, was used to obtain spectral amplitudes associated with EDA. We found that the time-varying spectral frequency band 0.08-0.24 Hz was most responsive to stimulation. Spectral power for frequencies higher than 0.24 Hz were determined to be not related to the sympathetic dynamics because they comprised less than 5% of the total power. The mean value of time-varying spectral amplitudes in the frequency band 0.08-0.24 Hz were used as the index of sympathetic tone, termed TVSymp. TVSymp was found to be overall the most sensitive to the stimuli, as evidenced by a low coefficient of variation (0.54), and higher consistency (intra-class correlation, 0.96) and sensitivity (Youden's index > 0.75), area under the receiver operating characteristic (ROC) curve (>0.8, accuracy > 0.88) compared with time-domain and time-invariant spectral indices, including heart rate variability.

Power Spectral Density Analysis of Electrodermal Activity for Sympathetic Function Assessment.

Time-domain features of electrodermal activity (EDA), the measurable changes in conductance at the skin surface, are typically used to assess overall activation of the sympathetic system. These time domain features, the skin conductance level (SCL) and the nonspecific skin conductance responses (NS.SCRs), are consistently elevated with sympathetic nervous arousal, but highly variable between subjects. A novel frequency-domain approach to quantify sympathetic function using the power spectral density (PSD) of EDA is proposed. This analysis was used to examine if some of the induced stimuli invoke the sympathetic nervous system's dynamics which can be discernible as a large spectral peak, conjectured to be present in the low frequency band. The resulting indices were compared to the power of low-frequency components of heart rate variability (HRVLF) time series, as well as to time-domain features of EDA. Twelve healthy subjects were subjected to orthostatic, physical and cognitive stress, to test these techniques. We found that the increase in the spectral powers of the EDA was largely confined to 0.045-0.15 Hz, which is in the prescribed band for HRVLF. These low frequency components are known to be, in part, influenced by the sympathetic nervous dynamics. However, we found an additional 5-10% of the spectral power in the frequency range of 0.15-0.25 Hz with all three stimuli. Thus, dynamics of the normalized sympathetic component of the EDA, termed EDASympn, are represented in the frequency band 0.045-0.25 Hz; only a small amount of spectral power is present in frequencies higher than 0.25 Hz. Our results showed that the time-domain indices (the SCL and NS.SCRs), and EDASympn, exhibited significant increases under orthostatic, physical, and cognitive stress. However, EDASympn was more responsive than the SCL and NS.SCRs to the cold pressor stimulus, while the latter two were more sensitive to the postural and Stroop tests. Additionally, EDASympn exhibited an acceptable degree of consistency and a lower coefficient of variation compared to the time-domain features. Therefore, PSD analysis of EDA is a promising technique for sympathetic function assessment.