BNNI2015 Summer School in Neuroinformatics

Multiscale computational neuroscience:

Neurons, networks and systems

Tartu, Estonia, 15.-18.6.2015

Introduction to Neuroinformatics

Marja-Leena Linne

Tampere University of Technology, Finland

INCF National Node of Finland

marja-leena.linne@tut.fi

Affiliated with:

European Institute of Theoretical Neuroscience, Paris, France

BioMediTech, University of Tampere, Finland

• Big questions and challenges in brain research

• Neuroinformatics and how it can help neuroscience

• International Neuroinformatics Coordination Facility

• Future application areas of neuroinformatics

• Other large-scale brain research initiatives

Contents

Big Questions in Brain Research

How does brain function?

What are the genetic, molecular, cellular, and network level

mechanisms that make the brain to

learn

remember

understand?

How to cure brain disorders and diseases?

Asperger syndrome

Autism

Anorexia

Mood disorders

Epilepsy

Parkinson’s disease

Alzheimer’s disease

Big Questions in Brain Research

Neurodegenerative diseases

• Alzheimer’s disease

• Parkinson’s disease

• Huntinton’s disease

• Common to all:

progressive loss of

structure or function

of neurons, including

death of neurons.

http://www.modernmedicalguide.com/alzheimers-disease/

Structure of the Human Brain

Cerebral Cortex: Structure and Function

Layers of cerebral cortex

Cerebellar Cortex: Structure and Function

Layers of cerebellar cortex

Frackowiak and Markram 2015

Spatial scaleTime scale

Str

uctu

re

Function

Molecular level

Cellular level

Tissue level

Whole

brain level

International Neuroinformatics Coordination Facility

Measurement Techniques – Lots of Data!

Neuroscience is increasingly fragmented

Results are compared in discussion sections,

but not quantitavely!

Different research groups use:

Different methods

Different brain areas

Different animals

Different levels of description

Future neuroscience

Challenges:

• To integrate knowledge.

• To link data across different levels of brain organization.

• To transfer knowledge from animals to humans.

• To understand the underlying mechanisms of brain diseases.

Neuroscience has to become “big science” and use “big data”.

NEUROINFORMATICS

Goal of Neuroinformatics

Integration of information across all levels and scales of

nervous system – from genes to behaviour.

Neuroinformatics is a newly emerged field that stands at the

intersection of neural science and information science.

INCF infrastructure will help facilitate the acquisition and annotation of diverse datasets

to achieve multiscale and multiomic data integration – from genes to behavior.

INCF ̶ International Neuroinformatics Coordination

Facility

Mission of INCF

• Coordinate and foster international activities in neuroinformatics

• Contribute to development and maintenance of database and

computational infrastructure and support mechanisms for

neuroscience applications

• Enable access to all freely accessible data and analysis resources

for human brain research to the international research community

• Develop mechanisms for the seamless flow of information and

knowledge between academia, private enterprises and the

publication industry

Training is one of the four strategic objectives

“Serve as a globally recognized leader for

the development and provision of training

and educational resources in

neuroinformatics”

INCF Strategy for 2016-20: Training

• Established in 2010 after three preparatory workshops, which

produced reports on Needs for Training in Neuroinformatics:

http://www.incf.org/activities/teaching-and-training

• Members (8):

Chair: David Willshaw (UK)

Gary Egan (Australia)

Gaute Einevoll (Norway)

Arvind Kumar (Germany/Sweden)

Marja-Leena Linne (Finland)

Barry Richmond (USA)

Thomas Wachtler (Germany)

Mathew Abrams (INCF Sekreteriat)

INCF Training Committee

1. Organizes 2-day Introduction to Neuroinformatics course

2. Part-funds international neuroinformatics-related courses

2011-14: 10 courses, 242 participants (€180.000)

3. Grants travel stipends for training

International NI courses

Visits to learn NI tools

4. Maintains online repository of neuroinformatics-related lectures, course materials, and tool demonstrations

videos at: https://www.youtube.com/user/INCForg/videos

5. Initiated INCFs involvement as mentoring organization in Google Summer of Code

A global program that offers students stipends to write code for open source projects

2011: 2, 2012: 6; 2013: 8; 2014: 13; 2015: 15.

Activities of the INCF Training Committee

Munich 2012

Stockholm 2013

Leiden 2014

Cairns 2015Student

75%

Post doc15%

Faculty6%

Other4%

Total of 120 participants (54%)

received financial support (€18.872)

Short Course: Introduction to Neuroinformatics

Subfields of Neuroinformatics

1.Neurobiology (fundamental, clinical)

2.Databases

3.Data analysis

a. Basic and advanced signal analysis, image

analysis

b. Visualization

4.Modeling (theoretical, data‐driven)

5.Simulation and computation (workflows)

6.Neuromorphic and neural engineering

What is Neuroinformatics in Practise?

Develops standards and “good practises” for

Data acquisition and storage

Provenance (ownership and usage)

Sharing

Publishing

Analysis

Visualization

Modeling and simulation

Develops federated (multiple into single)

databases

Develops theoretical, computational, and

simulation tools for quantitative modeling

Develop

methods to

ensure

interoperability

Develop

standardized

and

comprehensive

metadata

schemes

Are you willing to take

the time to prepare

your data for sharing?

Is there somewhere for

you to put your data?

Will it be possible to still

understand in the future?

Can I read your data?

If so, can I understand it?

(e.g., how it was acquired and

analysed?)

Develop

ontology and

automated

methods for

metadata

capture

Typical Situation in Neuroscience Research…

Develop

federated

database

systems

INCF Database

Collaboration between researchers through the storage, sharing, and

annotation of neuroscience data, text, images, sounds, movies,

models and simulations.

Database architecture

Databases and Ontologies: Challenges

How to make scientists to accept the idea of data sharing?

How to create citation system for experimental (and simulated) data?

How to present metadata of an experiment in an understandable,

machine-readable way?

How to combine multiple autonomous database systems into a single

federated (or virtual) database?

Ranjan et al., Front Neuroinform 2011Grewe et al., Front Neuroinform 2010

Computational neuroscience and

neuroinformatics

27

28

ResponseStimulus

Computational neuroscience is the field of study in

which mathematical tools and theories as well as

computer simulations are used to investigate brain

function.

It is an interdisciplinary science that links the diverse

fields of neuroscience, cognitive science, and

psychology with electrical engineering, computer

science, mathematics, and physics.

Top-down and Bottom-up models!

Multiscale Computational neuroscience

• Integrate experimental data

• Describe neural systems

• Fit existing data

• Make predictions about the system behavior in new conditions

• Provide mechanistic understanding of the neural system

• Provide new ways to study brain diseases

• Provide principles to develop new technology (neuromorphic systems, control for robots)

How Do Computational Models Help?

Applications:

From Integration of Data and Models to Clinical Neuroscience

Electrical high-frequency deep brain stimulation with humans:

Computational models are used for unlearning pathological neuronal

synchronization and synaptic coupling in Parkinson’s disease.

Peter Tass and colleagues

(Julich, Germany)

Measured brain

signalSimulation of desired

local activity

Comparison

(measured vs.

desired)

Applications:

From Integration of Data and Models to Clinical Neuroscience

Work on Aplysia:

Computational models are used for prediction of training (stimulation) protocols for

impaired function of molecular networks associated with learning and memory

Rong-Yu Liu, Paul Smolen, John Byrne and colleagues

(Texas Medical School in Houston)

Patient-tuned stimulation therapy ?

Applications:

Neuromorphic Engineering and Neurotechnology

How to create better technologies?

Intelligent toys

Intelligent devices

Autonomous cars

Robots

Neuroprostheses

Artificial Brain?

33

Australian Brain Project

Japanese Brain ProjectChinese Brain Project

Other emerging brain projects

Large-Scale Brain Initiatives

• Founder: Initiated in the US by non-profit organization.

• Aim: One Mind is a new-model non-profit organization that will

take the lead role in the research, funding, marketing, and public

awareness of mental illness and brain injury, by bringing together

the governmental, corporate, scientific, and philanthropic

communities in a concerted effort to drastically reduce the social

and economic effects of mental illness and brain injury within ten

years. With ONE MIND as the hub for open-source big data and

research, compelling the scientific, health care, and

pharmaceutical communities to collaborate completely for the

common good, we can decrease the time from research to

treatment/cure from 30 years to less than a decade.

34

ONE MIND

• Founders/Funding: Paul and Jody Allen

• Aim: to accelerate the understanding of how the human brain works in

health and disease. Using a big science approach, we generate useful

public resources, drive technological and analytical advances, and

discover fundamental brain properties through integration of

experiments, modeling and theory.

•

35

Allen Brain Institute

• The Human Brain-Project is a 10-year European initiative to launch a global, collaborative effort to understand the human brain.

• Officially launched 1 October, 2013

• 112 partner institutions, 24 countries

• Funding:

• €74M for the first 30 months

• €1 billion for 10 years

EU FET Flagship Human Brain Project

• To create the ICT infrastructure for

neuroscience and brain-related

research in medicine and computing.

• To build and simulate multiscale brain

models at different levels of detail.

The Goal of the Human Brain Project

38

SP1 - Strategic Mouse Brain Data

SP2 - Strategic Human Brain Data

SP3 - Cognitive Architectures

SP4 - Theoretical Neuroscience

SP5 – The Neuroinformatics Platform

SP6 – The Brain Simulation Platform

SP7 – The High Performance Computing Platform

SP8 – The Medical Informatics Platform

SP9 – The Neuromorphic Computing Platform

SP10 – The Neurorobotics Platform

SP11 – Applications

SP12 - Ethics and Society

SP13 - Management

EXPERIMENTAL DATA

MODELS OF NEURAL

SYSTEMS

(LINK BETWEEN DATA

AND TOOLS)

INFORMATICS TOOLS

FOR NEUROSCIENCE

Organization of the HBP: Subprojects

Neuroinformatics aims at bringing

all data measured from the brain

publicly available to help understand

brain functions in health and disease

Data analyses

Data visualization

Publishing

Mathematical

modeling

Data sharing

Computer

simulation

Education

and Training

Searching data

and information

Repeating experiments

Presenting new questions

and hypotheses

Summary