Signal Analysis for Neurocritical Patients

Analysis of Physiological Signals From Neurocritical Patients in Intensive Care Units Using Wavelet Transform and Deep Learning

The project uses big data analysis techniques such as wavelet transform and deep learning to analyze physiological signals from neurocritical patients and build a model to evaluate intracranial condition and to predict neurological outcome. By identification of correlations among these parameters and their trends, we may achieve early detection of anomalies and enhance the ability in judgement of current neurological condition and prediction of prognosis. By continuous input of the past and contemporary data in the ICU, the model will be modified repeatedly and its accuracy improves as the model grows. The model can be used to recognize abnormalities earlier and provide a warning system. Clinicians taking care of neurocritical patients can adjust their treatment policy and evaluate the outcome according to such system.

Study Overview

• Status: Completed
• Conditions: Brain Injuries, Acute
• Intervention / Treatment: Device: intracranial pressure monitoring

• Study Type: Observational
• Enrollment (Actual): 156

Contacts and Locations

Study Locations

• Taiwan
  • New Taipei City, Taiwan, 200
    • Far Eastern Memorial Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 20 years and older (ADULT, OLDER_ADULT)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

Neurocritical patients admitted to intensive care unit (ICU), including but not limited to traumatic brain injury, hemorrhagic stroke, ischemic stroke, brain infection, brain tumor and acute hydrocephalus.

Description

Inclusion Criteria:

• Age equal to or older than 20 years
• Neurocritical patients admitted to intensive care unit (ICU), including but not limited to traumatic brain injury, hemorrhagic stroke, ischemic stroke, brain infection, brain tumor and acute hydrocephalus.
• Patients who have undergone cranial surgery and had intracranial pressure monitor inserted or external ventricular drainage. The central monitor of ICU is able to collect the data continuously

Exclusion Criteria:

• Age younger than 20 years.
• Continuous monitoring of intracranial pressure is not feasible.

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Neurocritical patients; Patients with brain injury from trauma, ischemic stroke, hemorrhage stroke (intracerebral hemorrhage, subarachnoid hemorrhage), brain tumor with increased intracranial pressure, brain infection, hydrocephalus, among others.	Device: intracranial pressure monitoring; The patients may have either intracranial pressure (ICP) monitor insertion or external ventricular drainage that can be used as ICP monitor.; Other Names: physiological monitoring

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Neurological status	Glasgow coma scale/Mortality	Discharge out of the intensive care unit, averaged 2 weeks

Collaborators and Investigators

• Sponsor: Far Eastern Memorial Hospital
• Investigators: Principal Investigator: Yi-Hsin Tsai, M.D., Far Eastern Memorial Hospital

Publications and helpful links

General Publications

• Diederik P. Kingma and Jimmy Lei Ba. Adam: A method for stochastic optimization. Conference paper at ICLR 2015.
• Michael Unser and Akram Aldroubi. (1996 Apr) A review of wavelets in biomedical applications. Proceedings of the IEEE 84(4): 626-638.
• LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015 May 28;521(7553):436-44. doi: 10.1038/nature14539.
• Christopher Torrence and Gilbert P. Compo. (1998 Jan) A practical guide to wavelet analysis. Bulletin of the American Meteorological Society 79(1):61-78.
• Theis, Fabian & Meyer-Base, Anke. (2010). Biomedical Signal Analysis - Contemporary Methods and Applications. Biomedical Signal Analysis: Contemporary Methods and Applications.
• Min S, Lee B, Yoon S. Deep learning in bioinformatics. Brief Bioinform. 2017 Sep 1;18(5):851-869. doi: 10.1093/bib/bbw068.
• Yi Mao, Wenlin Chen, Yixin Chen, Chenyang Lu, Marin Kollef, and Thomas Bailey. (2012) An integrated data mining approach to real-time clinical monitoring and deterioration warning. Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining Pages 1140-1148. doi>10.1145/2339530.2339709

Study record dates

Study Major Dates

• Study Start (ACTUAL): December 18, 2017
• Primary Completion (ACTUAL): May 24, 2018
• Study Completion (ACTUAL): May 31, 2018

Study Registration Dates

• First Submitted: November 29, 2017
• First Submitted That Met QC Criteria: December 4, 2017
• First Posted (ACTUAL): December 5, 2017

Study Record Updates

• Last Update Posted (ACTUAL): June 6, 2018
• Last Update Submitted That Met QC Criteria: June 3, 2018
• Last Verified: June 1, 2018

More Information

Terms related to this study

• Keywords: Deep learning; Wavelet transform
• Additional Relevant MeSH Terms: Brain Diseases; Central Nervous System Diseases; Nervous System Diseases; Wounds and Injuries; Craniocerebral Trauma; Trauma, Nervous System; Brain Injuries
• Other Study ID Numbers: 106152-E

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No