Change search
ReferencesLink to record
Permanent link

Direct link
Stroke Damage Detection Using Classification Trees on Electrical Bioimpedance Cerebral Spectroscopy Measurements
KTH, School of Technology and Health (STH), Medical Engineering, Medical sensors, signals and systems.ORCID iD: 0000-0002-0928-8501
KTH, School of Technology and Health (STH), Medical Engineering, Medical sensors, signals and systems.ORCID iD: 0000-0002-6995-967X
KTH, School of Technology and Health (STH), Medical Engineering, Medical sensors, signals and systems.
2013 (English)In: Sensors, ISSN 1424-8220, Vol. 13, no 8, 10074-10086 p.Article in journal (Refereed) Published
Abstract [en]

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.

Place, publisher, year, edition, pages
2013. Vol. 13, no 8, 10074-10086 p.
Keyword [en]
stroke, electrical bioimpedance spectroscopy, classification tree, cole parameters
National Category
Medical Engineering
URN: urn:nbn:se:kth:diva-140393DOI: 10.3390/s130810074ISI: 000328624800027PubMedID: 23966181ScopusID: 2-s2.0-84922971548OAI: diva2:690257

QC 20140123

Available from: 2014-01-23 Created: 2014-01-23 Last updated: 2015-11-09Bibliographically approved
In thesis
1. Electrical Bioimpedance Cerebral Monitoring: From Hyopthesis and Simulation to First Experimental Evidence in Stroke Patients
Open this publication in new window or tab >>Electrical Bioimpedance Cerebral Monitoring: From Hyopthesis and Simulation to First Experimental Evidence in Stroke Patients
2015 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Stroke is amongthe leading causes of death worldwide and requires immediate care to prevent death or permanent disability. Unfortunately, the current stateof stroke diagnosis is limited to fixed neuroimaging facilities that do not allow rapid stroke diagnosis. Hence, a portable stroke-diagnosis device could assist in the pre-hospital triage of patients. Moreover, such a portable device could also be useful for bedside stroke monitoring of patients in the Neuro Intensive Care Unit (Neuro-ICU) to avoid unnecessary neuroimaging. Recent animal studies and numerical simulations have supported the idea of implementing Electrical Bioimpedance (EBI) in a portable device, allowing non-invasive assessment as a useful tool for the pre-hospital triage of stroke and Traumatic Brain Injury (TBI) patients. Unfortunately, these studies have not reported any results from human subjects in the acute phase of the stroke. The numerical simulations are also based on simple models that sometimes lack necessary details.

Finite Element Method (FEM) simulations on a realistic numerical head model as well as experimental Bioimpedance Spectroscopy (BIS) measurements from human subjectsin the acute, subacute and chronic phasesof stroke were used to answer the following research questions: (i) Does stroke modify the electrical properties of brain tissue in a way that is detectable via EBI? (ii) Would it be possible to detect stroke via EBI as early as in the acute and sub-acute phase?(iii) Is EBI sensitive enough to monitor changes caused by stroke pathogenesis?

Using FEM to simulate electrical current injection on the head and study the resulting distribution of electrical potential on the scalp, it was shown that Intra-Cranial Hemorrhage (ICH) affects the quasi-symmetric scalp potential distribution,creating larger left-right potential asymmetry when compared to the healthy head model. Proof-of-concept FEM simulations were also tested in a small cohort of 6ICH patients and 10 healthy controls, showing that the left-right potential difference in the patients is significantly (p<0.05) larger than in the controls. Using bioimpedance measurements in the acute,  subacute and chronic phasesof stroke and examining simple features, it was also shown that the head EBI measurements of patients suffering stroke are different from controls, enabling the discrimination of healthy controls and stroke patients at any stage of the stroke. The absolute change in test-retest resistance measurements of the control group (~5.33%) was also found to be significantly (p<0.05) smaller than the EBI measurements of patients obtained 24 hours and 72 hours after stroke onset (20.44%). These results suggested that scalp EBI is sensitive to stroke pathogenesis changesand thususeful for bedside monitoring in the Neuro-ICU. These results suggested that EBI is a potentially useful tool for stroke diagnosis and monitoring.

Finally, the initial observations based on a small number of patients, addressing the proposed future work of this thesis, suggested that the average head resistance amplitude of hemorrhagic stroke patients is smaller than in healthy controls, while ischemic stroke patients show a larger resistance amplitude than the controls. Scalp potential asymmetry analysis of healthy, hemorrhagic and ischemic stroke subjects also suggests that these three groups can be separated. However, these results are based on a small number of patients and need to be validated using a larger cohort. Initial observations also showed that the resistance of the EBI measurements of controls is robust between test and retest measurements, showing no significant difference (less than 2% and p>0.05). Subject position during EBI recording (supine or sitting) did not seem to affect the resistance of the EBI measurements (p>0.05). However, age, sex and head size showed significant effects on the resistance measurements. These initial observations are encouraging for further research on EBI for cerebral monitoring and stroke diagnosis. However, at this stage, considering the uncertainties in stroke type differentiation, EBI cannot replace CT but has the potential to be used as a consultation tool.

Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology, 2015. x, 72 p.
TRITA-STH, 2015:8
Electrical Biompedance Spectroscopy, Stroke, Hemorrhage, Ischemia, FEM, HFSS, Electrical Potentials
National Category
Medical Engineering
urn:nbn:se:kth:diva-176634 (URN)978-91-7595-769-2 (ISBN)
Public defence
2015-12-01, Sal 3-221, Alfred Nobels Allé 10 141 52, Huddinge, 13:15 (English)

QC 20151109

Available from: 2015-11-09 Created: 2015-11-09 Last updated: 2015-11-09Bibliographically approved

Open Access in DiVA

No full text

Other links

Publisher's full textPubMedScopus

Search in DiVA

By author/editor
Atefi, Seyed RezaSeoane, FernandoLindecrantz, Kaj
By organisation
Medical sensors, signals and systems
In the same journal
Medical Engineering

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 172 hits
ReferencesLink to record
Permanent link

Direct link