Stroke Damage Detection Using Classification Trees on Electrical Bioimpedance Cerebral Spectroscopy Measurements

Abstract

After cancer and cardio-vascular disease, stroke is the third greatest cause of death worldwide. Given the limitations of the current imaging technologies used for stroke diagnosis, the need for portable non-invasive and less expensive diagnostic tools is crucial. Previous studies have suggested that electrical bioimpedance (EBI) measurements from the head might contain useful clinical information related to changes produced in the cerebral tissue after the onset of stroke. In this study, we recorded 720 EBI Spectroscopy (EBIS) measurements from two different head regions of 18 hemispheres of nine subjects. Three of these subjects had suffered a unilateral haemorrhagic stroke. A number of features based on structural and intrinsic frequency-dependent properties of the cerebral tissue were extracted. These features were then fed into a classification tree. The results show that a full classification of damaged and undamaged cerebral tissue was achieved after three hierarchical classification steps. Lastly, the performance of the classification tree was assessed using Leave-One-Out Cross Validation (LOO-CV). Despite the fact that the results of this study are limited to a small database, and the observations obtained must be verified further with a larger cohort of patients, these findings confirm that EBI measurements contain useful information for assessing on the health of brain tissue after stroke and supports the hypothesis that classification features based on Cole parameters, spectral information and the geometry of EBIS measurements are useful to differentiate between healthy and stroke damaged brain tissue.

Intracranial hemorrhage alters scalp potential distribution in bioimpedance cerebral monitoring: Preliminary results from FEM simulation on a realistic head model and human subjects

“…Clinical results specifically confirmed that there exists a statistically significant (P < 0.05) larger potential difference in the scalp potential distribution of patients suffering acute/subacute ICH than the controls, which, unlike simulated results, can be measured relatively fast (<1 h). Although these results are in agreement with the results from other numerical and experimental studies, 32,34,35,37,45,46 unlike previous experimental studies using stroke lesions from the chronic phase, i.e., lesions over a month old, 34,35,45 we obtained EIS recordings from patients in the acute and subacute phases, i.e., lesions less than five days old, to demonstrate the potential usefulness of EBI for the prehospital triage of patients suffering acute/subacute brain injury.…”

“…Despite the limitations of the study, like the low statistical power, the obtained results agree with previous bioimpedance-related studies. 32,34,35,37,45,46 The results indicate that a hemorrhagic brain lesion generates spatial changes in voltage, which can be recorded directly on scalp, thus supporting this type of bioimpedance technology for potential early detection of brain ICH.…”

“…31 In the past decade, EBI has also been proposed as a potentially useful technique for stroke detection. [32][33][34][35] With regard to ICH, in vitro EIS and EIT studies suggest that hemorrhage in the brain will modify the tissue's electrical properties, as the conductivity of blood is about three times greater than that of brain tissue. 36 Unfortunately, previous quantitative studies have been based on simple models and not supported by human bioimpedance measurements.…”

“…32,37 Experimental studies have also been limited to the chronic phase of the stroke that is not relevant to the ultimate goal of bioimpedance technology for detecting acute/subacute ICH. 34 In this study, 3D accurate finite elements method (FEM) …”

A comparative study of PCA, SIMCA and Cole model for classification of bioimpedance spectroscopy measurements

“…Simplicity, popularity and explanatory power of these features make them the first choice and the standard way of descriptive feature extraction, as well. However, researchers found that Cole parameters extracted from repetitive measurements on the same subject are too variable and thus not effective in achieving a classification task [14,15].…”