Tutorials
The role of the tutorials is to provide a platform for a more intensive scientific exchange amongst researchers interested in a particular topic and as a meeting point for the community. Tutorials complement the depth-oriented technical sessions by providing participants with broad overviews of emerging fields. A tutorial can be scheduled for 1.5 or 3 hours.
New directions for biosignal processing : BCIs for coma and stroke patients
Instructors
|
Brendan Allison
Independent Researcher
Austria
|
|
|
|
Brief Bio
Brendan Allison earned his PhD in Cognitive Science in 2003 from the University of California at San Diego. He has been active in BCI Research for about 20 years, including postdoctoral work for Profs. Jonathan Wolpaw, Gert Pfurtscheller, and Christa Neuper. He recently joined Guger Technologies as a senior scientist.
|
|
Slav Dimov
Sales,
Austria
|
|
|
|
Brief Bio
Slav Dimov earned his Master in Economics in 2009 and Bachelor in Microsystems Engineering in 2015 from the University of Freiburg, Germany. He recently joined Guger Technologies as a sales officer.
|
Abstract
Abstract:
One of the most exciting applications of biosignal processing is brain-computer Interfaces (BCIs), in which users can spell or perform other tasks via thought alone. Very recent work has extended BCI Technology to help persons with coma and stroke. We will provide interactive, hands-on demonstrations of new BCI technologies.
Keywords:
BCI, Brain-Computer Interface, BMI, stroke, coma, EEG
Aims and Learning Objectives:
Attendees will learn about Signal processing methods used in BCIs and how they can be extended to provide real-world help for patients.
Target Audience:
engineers, mathematicians, signal processing experts, students
Detailed Outline:
The first 30 minutes will present a general introduction to BCIs. We will then provide a 30 Minute lecture with details of signal processing approaches and relevance to stroke and coma patients. The remaining two hours will be devoted to Hands-on interactive demos.
Keywords
BCI, Brain-Computer Interface, BMI, stroke, coma, EEG
Aims and Learning Objectives
Attendees will learn about Signal processing methods used in BCIs and how they can be extended to provide real-world help for patients.
Target Audience
engineers, mathematicians, signal processing experts, students
Prerequisite Knowledge of Audience
None.
Detailed Outline
The first 30 minutes will present a general introduction to BCIs. We will then provide a 30 Minute lecture with details of signal processing approaches and relevance to stroke and coma patients. The remaining two hours will be devoted to Hands-on interactive demos.
Identification of Signatures and Classification – Application to the Systems Toxicology Computational Challenge
Instructor
|
Vincenzo Belcastro
Independent Researcher
Switzerland
|
|
|
|
Brief Bio
Vincenzo graduated in Computer Science at the University of Naples in 2007 with a dissertation on statistical and mathematical approached to reverse-engineer gene network from gene expression data. Awarded of a PhD from the Open University of Cambridge, he developed a mutual information based algorithm to infer mammalian gene networks from a massive gene expression datasets. The project was developed at the Telethon Institute of Genetic and Medicine (TIGEM) in Naples, with specific emphasis to disease genes. After his PhD, Dr. Belcastro moved to Philip Morris International (PMI R&D), where he developed and applied computational approaches to study the impact of smoke on both in vivo and in vitro biological systems. Since the very first months, Dr Belcastro was involved in an innovative research project jointly run by PMI and IBM, the Systems Biology Verification (SBV), IMPROVER. Dr Belcastro actively contributed to the organization of the Species Translation challenge, and to the set-up of the Network Verification challenge (https://www.sbvimprover.com/). These activities eventually resulted in the publication of scientific manuscripts in leading journals. Dr Belcastro is currently co-organizer of the Systems Toxicology Challenge (SysTox), whose objective is to verify that robust and sparse human specific-independent gene signatures of exposure response can be extracted in whole blood gene expression data from human, or human and rodent to predict exposed and non-exposed group labels.
|
Abstract
Abstract:
Personalized medicine and risk assessment in the context of 21st century toxicology rely on the identification of signatures that allow to predict for example specificities of the disease, what would be the best course of treatment or to identify specific exposure markers to elucidate mechanisms of toxicity. Gene signatures have been extracted and applied more or less successfully for decades now. Interestingly, there are still many definitions of what gene signatures entail. As a foundation of predictive medicine and risk assessment, datasets are generated using high-throughput technologies to test diverse set of chemicals or mixtures in different biological systems. However, the development of effective computational approaches for the analysis and integration of these data sets remains challenging. The sbv IMPROVER (Industrial Methodology for Process Verification in Research; http://sbvimprover.com/) project aims to verify methods and concepts in systems biology research via challenges open to the scientific community. In 2012, the diagnostic signature challenge allowed participants to benchmark methods for classification of transcriptomics profiles from patients suffering of diverse diseases. A perennial benchmarking site is available online for scientists to continue benchmarking their methods using the datasets. In fall 2015, the 4th sbv IMPROVER computational challenge was launched to evaluate algorithms for the identification of specific markers of chemical mixture exposure response in blood of humans or rodents. This computational challenge addresses questions related to the classification of samples based on transcriptomics profiles from well-defined sample cohorts and its results should be generalizable to other fields benefiting from sample classification and prediction methods. Moreover, it will address whether gene expression data derived from human or rodent whole blood are sufficiently informative to identify human-specific or species-independent blood gene signatures predictive of the exposure status of a subject to chemical mixtures (current/former/non-exposure). The proposed tutorial will review the history and methods used for signature extraction and sample classification, focusing primarily in machine learning methods. It will allow participants to discover the benchmarking platform to train their methods, and possibly apply them later to take part in the open computational challenge related to exposure response.
Keywords:
Signatures; Machine learning; Computational challenge; Risk assessment; Exposure response markers.
Aims and Learning Objectives:
This tutorials aims are 1) introducing the concepts and methods relevant to signature extraction and sample classification from big data; 2) presenting a real industrial/pharma scenario with a specific biological question of interest (exposure marker discovery); and 3) provide guidelines on how to address the challenge by applying computational and machine learning technics.
Target Audience:
Persons with a background in any of the following area: computational biology; systems biology; bioinformatics; mathematics; statistics; machine learning
Detailed Outline:
60 minutes – Introduction to the concepts of signature and classification using machine learning. Short description of past biological challenges addressed via computational technics in the sbv IMPROVER project. Outcome and lesson learned from the crowdsourcing initiative. 20 minutes – Presentation of the diagnostic signature challenge and the benchmarking tool available on the sbv IMPROVER platform, as well as concepts relevant to the open computational challenge. 60 minutes – Hands on session. Participants will be asked to register to the web platform, download data, and apply machine learning technic to solve the proposed challenge (data loading scripts will be provided to facilitate data manipulation). They will be able to get live scoring of their performance. 40 minutes – Review results and open discussion.
Keywords
Signatures; Machine learning; Computational challenge; Risk assessment; Exposure response markers.
Aims and Learning Objectives
This tutorials aims are 1) introducing the concepts and methods relevant to signature extraction and sample classification from big data; 2) presenting a real industrial/pharma scenario with a specific biological question of interest (exposure marker discovery); and 3) provide guidelines on how to address the challenge by applying computational and machine learning technics.
Target Audience
Persons with a background in any of the following area: computational biology; systems biology; bioinformatics; mathematics; statistics; machine learning.
Prerequisite Knowledge of Audience
Familiarity with any programming environment. The R environment will be used for study cases.
Detailed Outline
60 minutes – Introduction to the concepts of signature and classification using machine learning. Short description of past biological challenges addressed via computational technics in the sbv IMPROVER project. Outcome and lesson learned from the crowdsourcing initiative.
20 minutes – Presentation of the diagnostic signature challenge and the benchmarking tool available on the sbv IMPROVER platform, as well as concepts relevant to the open computational challenge.
60 minutes – Hands on session. Participants will be asked to register to the web platform, download data, and apply machine learning technic to solve the proposed challenge (data loading scripts will be provided to facilitate data manipulation). They will be able to get live scoring of their performance.
40 minutes – Review results and open discussion.