Program

The Brainhack will start on Thursday December 2nd late afternoon with project pitches and teams creation. The following two days, Friday 3rd and Saturday 4th, will be full of hacking, but also breaks in the form of talks on cutting edge neuroimaging research

08:30
Breakfast
Breakfast
09:00
Ignite talk
Open Hacking
09:30
Open Hacking
Open Hacking
12:00
Lunch
Lunch
13:00
Ignite talk
Open Hacking
14:00
Open Hacking
Open Hacking
16:30
Collect Badges
Open Hacking
Open Hacking
17:00
Intro / Project pitches
Open Hacking
Debrief Projects
18:00
Choose your team!
Cross-team activity
Wrap up / Drinks
19:00
Social Drinks
Social Drinks / Dinner

Registration

Registration is FREE but MANDATORY.

Please click on the button below and fill out the form by December 1st, 2021.

Registration

Team

We are here to help. Don't hesitate to ask us any question.

Speakers

Participating to provide examples of cutting-edge neuroimaging research.



"Dealing with data and annotation scarcity in machine learning for clinical neuroimaging" by Jonas Richiardi

Machine learning (ML) algorithms perform better with increased dataset sizes, and clinical applications of ML to neuroimaging are no different. However, clinical imaging data, in particular retrospective data, is not intrisincally suitable for ML: obtaining homogeneous, complete, and suitably annotated images in sufficiently large numbers is difficult to achieve.

In this talk, I will first discuss issues of domain shift and the impact of protocol differences, and possible mitigation strategies including harmonisation. Then, I will show possible approaches to speeding up image annotations using a weak labelling approach, and finally show atlas-based approaches to diffusion imaging when no such data is available. The talk will focus on our work on stroke, dementia, aneurysms and multiple sclerosis undertaken at the Translational machine learning laboratory, Lausanne University Hospital and University of Lausanne (together with our collaboration partners), and include other efforts worldwide in these areas.

"An introduction to brain fingerprinting" by Enrico Amico

In the 17th century, physician Marcello Malpighi observed the existence of distinctive patterns of ridges and sweat glands on fingertips. This was a major breakthrough, and originated a long and continuing quest for ways to uniquely identify individuals based on fingerprints, a technique massively used until today. It is only in the past few years that technologies and methodologies have achieved high-quality measures of an individual’s brain to the extent that personality traits and behavior can be characterized.

The concept of “fingerprints of the brain” is very novel and has been boosted thanks to a seminal publication by Finn et al. in 2015. They were among the firsts to show that an individual’s functional brain connectivity profile is both unique and reliable, similarly to a fingerprint, and that it is possible to identify an individual among a large group of subjects solely on the basis of her or his connectivity profile.

Yet, the discovery of brain fingerprints opened up a plethora of new questions. In particular, what exactly is the information encoded in brain connectivity patterns that ultimately leads to correctly differentiating someone’s connectome from anybody else’s? In other words, what makes our brains unique?

In this talk I am going to partially address these open questions while keeping a personal viewpoint on the subject. I will outline the main findings, discuss potential issues, and propose future directions in the quest for identifiability of human brain networks.

"Psychedelic Research 2.0: Psychedelics in the age of neuroscience" by Cyril Petignat

Widely popularized in the 1960s under the impulse of personalities such as Aldous Huxley or Timothy Leary, psychedelic substances have been the object of promising scientific research in the psychiatric field. Their use by the general population and the counter-culture movement led Richard Nixon in 1971 to ban their use, both therapeutic and recreational. But in the early 2000s, research resumed, and psychedelics came back to the fore. The movement accelerates again recently with the creation of research centers, student associations and conferences on the subject.

In this presentation, L' Association pour la Recherche Psychédélique de l'Université de Genève (ARP) will have the pleasure to review the modern history of these particular substances. We will also present recent research combining modern imaging (fMRI, EEG, ...) and psychedelics drugs. Neuroscience and psychedelics are intimately linked since their neurobiological mechanism induces changes in brain activity, this process might help us to understand the psychedelic experience and different state of consciousness. We will also briefly explain how this substance can be used in a therapeutic context in order to improve mental health.

Projects

Projects can be filered by keywords / skills

  • All
  • Programming
  • Terminal / Git
  • Machine Learning
  • MRI processing
  • EEG processing
  • Virtual Reality
  • Neurosciences


You can even propose your own by filling the form here!

Project 1 Add a real-time annotation system compatible with MNE to mark bad segments of EEG data on BSL's real-time viewer using pyqtgraph and pyopengl by Mathieu Scheltienne

EEG data is noisy by nature, and MNE has a very nice feature called annotations to mark bad segments of data. Those segments are then rejected internally by functions accepting the argument reject_by_annotation.
Until recently, interactive annotation on the MNE browser was very painful as the default browser (up to version 1.0) uses an old matplotlib backend... as slow as it gets. But, with version 0.24, a new backend has been introduced using pyqtgraph and pyopengl to render on the GPU. And it is amazing!
But even with this new backend, annotating raw data takes a lot of time, especially on large scale study.
This is where this project comes in. During a typical EEG session, someone will be monitoring the signal. In itself, this is a lightweight task, that could be coupled with a system to annotate in real-time the monitored signal. No more annotation in a later post-processing stage!

The FCBG platform has develop a library called NeuroDecode which has a build-in real-time viewer. This library uses LSL to stream EEG data, and pyqtgraph (rings a bell?) for the viewer. All the data streaming/acquisition/viewing capabilities have been moved and improved in BSL.

Basically, this project doesn't start from scratch! You can find here an example of the real-time viewer that will be modified for this project. The viewer has 2 backends: pyqtgraph and vispy. The second is incomplete, but runs efficiently on GPU. However, it is a bit obsolete as new version of pyqtgraph are now compatible with pyopengl... to run efficiently on GPU.

The objective is to add pyopengl, improve the GUI/add an annotation section, and add click and drag annotation on the backend window.

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project 2 EPILEPTIC EVENTS CLASSIFICATION by FABIEN FRISCOURT

GOAL OF THE PROJECT

Epilepsy is a frequent neurological condition characterized by episodes of abnormal highly synchronized activity of a population of neurons called epileptic seizure. In the kainate model of temporal lobe epilepsy, the disease is induced by an injection of kainate in one hippocampus, leading to the development of spontaneous recurrent seizures (SRS) after a latent phase of 3-4 weeks. While SRS characterize this later so-called chronic phase, other types of epileptic pathological events, differing in duration, amplitude and/or frequency content from the seizures, already appear a few days after the induction of the epileptic condition. For a specialist eye, those events can be defined and categorized but this process is time consuming and most importantly it remains subjective to a certain extent. The goal of this Brainhack project is to develop a pipeline that will perform this pathological epileptic events detection and classification blindly and automatically.


SCIENTIFIC CONTEXT

Epilepsy is a neurological condition that affect around 50 million people worldwide (1% prevalence). This pathology is defined by recurrent and incapacitating seizures that are often associated with cognitive disorders. Different kind of seizures can be observed according to the onset localization:

  1. Focal seizures that affect a constraint brain region
  2. Secondary generalized seizure where the onset is well defined but from which epileptiform activities will propagate to distant brain regions
  3. Generalized seizure in which the onset is in both hemisphere

This brainhack project will take place in an animal model of focal epilepsies: the kainate model of temporal lobe epilepsy. The procedure is simple, 0.1µg of kainate is injected in the mouse hippocampus. This led to a non-convulsive generalized seizure called status epilepticus (SE) that last several hours. After this SE, a latent phase (±21-28 days) is observed is which no seizure can be seen. The chronic phase starts when spontaneous and recurrent seizures (SRS) appear.

Even though SRS occur in the chronic phase, other pathological activities can be observed before, during the latent phase. Those activities can be categorized by an expert, but the goal here is to develop a pipeline that can do it automatically and blindly. Visually, 4 types of pathological activities were observed (but it’s possible to have more than 4 types in reality).

(Will be described in details later)


HOW CAN YOU DO? 

Different ways can be used to categorized those events and see if there are more than 4 categories. First, we need to identified features easily usable and personalized. The goal of the project will be to apply machine learning approach to automatically detect those events: First step would be de find best features of the signal (amplitude, frequency power, others .. ?) that could help detect epileptics events.

One could then train surpervised machine learning algorithm to automatically classify events based on the labels in the datasets.

One could also try unsupervised algorithm to the define new cluster of events and see if they correspond to particular (epileptics) events in the recordings. This approach could help to define new type of epileptic events that are not visually recognizable.

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project 3 Misfolded proteins spreading by Aleksandra Pestka

Most of neurodegenerative diseases are believed to be related to the spreading of misfolded proteins. Controversies still arise whether which path those proteins follow, and whether brain structures are damaged during this spreading or they were predisposing it.

In the project, DWI data collected from Alzheimer's patients is used. The data flow includes: preprocessing, tractography generation, connectivity matrix calculation and connectivity graph visualization.
The next step is to simulate a spreading of misfolded proteins based on generated connectivity matrix and the location of seeds for Alzheimer's beta-amyloid in the brain.

Goals for Brainhack Geneva:
- discuss and code the simulation methods (heat kernel based diffusion, epidemic simulation model etc.)
- discuss the problem of a connectome structure change in time caused by neurodegeneration

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project 4 Can blockchain tackle (some) problems in scientific publishing and financing ? Ideals, Ideas & implementation by Victor Férat

Publication and funding are major issues in scientific research. However, these two processes are nowadays centralized around institutions and journals. Researchers who play a major role in these processes rarely have a voice in the decisions of these institutions. In recent years, blockchain has proven to be a powerful tool when it comes to solving centralization problems, DeFI (decentralized finance) being a good example. Whether you are a fan of this technology or a fervent opponent, it is important to explore and discuss the new applications that blockchain can bring to the scientific world, in order to make the most of it but also to limit the excesses that it could generate. For this project, we invite you to come and discuss, exchange and redefine the possible links between the different actors of the scientific life in the light of recent advances of the blockchain world. If we feel like it, we can work on developing a proof-of-concept of our ideas. No knowledge is required, just curiosity and a well-honed critical mind. Suggested reading: https://polkadot.network/blog/polkadot-governance/ https://arxiv.org/abs/2101.09378 https://www.statnews.com/2018/12/21/reinvent-scientific-publishing-blockchain

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project 5 Contributing to Connectome Mapper 3, an Open-Source processing pipeline software for mapping the human connectome at multiple scales by Sebastien Tourbier

Connectome Mapper 3 is an open-source pipeline software, released as a BIDS App, that provides flexible pipelines for both diffusion MRI and functional MRI for mapping hierarchical multi-scale connectomes from multi-modal datasets. During the brainhack, any kind of contribution to this open-source project is welcome ✨. For instance, you can:

  1. Make CMP3 adopting all contributors to manage and acknowledge any kind of contribution.
  2. Create tutorial notebooks reading and analysing connectome files.
  3. Join the task force that extends Connectome Mapper 3 to EEG. (See GitHub Pull Request)
  4. Start working on the generation of a processing summary report.
  5. Propose your own.

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project 6 IMaging-PsychiAtry Challenge (IMPAC): Predicting autism diagnosis from MRI data by Yuliia Nikolaenko

The aim of the project is to try to predict autism diagnosis from MRI data. The ABIDE public dataset will be used for trying and training algorithms, and the NYU site will be kept separate for scoring. We will compare the results with those obtained in the previous challenge we organised in 2018 (Traut et al. 2021).

Project skills

At least one of these preferred, but contributing to general ideas / discussion is always welcome!

Programming
Terminal / Git
Machine Learning
MRI processing
EEG processing
Virtual Reality
Neurosciences

Project instigators

Here is the list of project instigators who will help giving the "big picture" of each project.

Mathieu Scheltienne

EPFL - MIP Lab / FCBG

Mathieu's expertise: cneuroimaging, EEG, python, MNE-Python, mahine learning

FABIEN's expertise: Epilepsy, Animal model, sEEG

Aleksandra's expertise: Data Science, Machine Learning, Neuroimaging, Python

Victor's expertise: EEG, Python

Sebastien's expertise: Reproducible Analysis, Data Visualization, Software Development, C++/Python

Yuliia Nikolaenko

Institut Pasteur

Yuliia's expertise: Social Informatics, Digital Sciences, Machine Learning, NLP

Contact & Venue

Address

Chemin des Mines 9, 1202 Geneva, Switzerland

Phone Number

Email