Versión en español

The BrainCode Games

Description

The BrainCode Games is a hackathon, a kind of collaborative competition, which ultimate goal is to develop a solution to an open problem in a short period of time (but very much intense!). The idea of The BrainCode Games is to use Artificial Intelligence to solve a Neuroscience problem. Don’t worry if you feel lacking the required knowledge or skills. The final objective of this hackathon is that you learn, in a didactic, practical and fun way, about some hot topics of the field (artificial intelligence, neuroscience and programming). For this reason, we will organize a series of lectures before the event that you can follow remotely. You will receive an attendance certificate for training and participation hours. And there will be awards for the team with best results.

Organization

This event is organized by:

• Andrea Navas-Olive: Computational Neuroscience PhD student at Instituto Cajal - CSIC, and teacher in the Bachelor degree in Computer Science Engineering at Universidad Autónoma de Madrid (UAM).

• Rodrigo Amaducci: Computational Neuroscience PhD student and teacher in the Bachelor degree in Computer Science Engineering at Universidad Autónoma de Madrid (UAM).

• Enrique R Sebastian: Computational Neuroscience PhD student at Instituto Cajal - CSIC.

• Daniel García-Rincón: PhD in Biochemistry, Molecular Biology and Biomedicine by Universidad Complutense de Madrid, and currently Data Scientist at BBVA Advanced Analytics at Indizen.

• Adrián Gollerizo: Physics and Technology teacher at IDEO Escuela and Education PhD student at Universidad Autónoma de Madrid (UAM).

• Liset M de la Prida: Director of the Hippocircuit Lab from Instituto Cajal - CSIC.

Lectures and material

• Neuroscience: For people with more technical backgrounds, we will review the mechanisms that our brain has to process information. Starting from the simplest elements, the neurons, up to neural circuits, and how they build a neural representation of the world. We will study what are neural oscillations and how they are related to behaviour. Conducted by Enrique R Sebastian.

• Python: For those with little programming experience, or with experience other programming languages, there will be an introductory course to Python, covering from the most simple and usual syntax to small programs using artificial intelligence tools that will be useful in the hacktahon. Conducted by Andrea Navas-Olive.

  • Material and exercises:
    1. Tutorial: Introduction
    2. Tutorial: Control Flow Tools
    3. Tutorial: Data Structures
    4. Tutorial: Modules

• Artificial Intelligence: we will explain what is Artificial Intelligence and Machine Learning. We will cover various strategies to target different problems, with a special focus on supervised-learning algorithms, since they better fit our open problem. Conducted by Daniel García-Rincón.

  • Material and exercises:
    1. Lineae Regression
    2. Logistic Regression
    3. KNN
    4. SVM
    5. k-Means
    6. Fraud Detection (for this exercise you will need this dataset)

Important information

• Hackathon dates: the weekend from 22th to 24th of October of 2021.
• There will be training lectures before the event. These will be recorded and uploaded to the website so you can access them at any moment.
• Both the hackathon and the lectures will be held online.
• We will open online platforms to meet and discuss during the hackathon.
• Limited slots available.
• Students will be prioritized.
• Programming language: Python.
• Awards: up to 500€ for the team with best results.
• Attendance certificate will be provided.

More about…

See below for more information. If you have any question, do not hesitate in contact us: braincodegames@gmail.com

…the open problem

The problem consists in detecting certain events from electrophysiological signals (which are recordings of the brain activity). These events, called sharp wave ripples or just ripples, are a kind of fast oscillations as those shown in the picture below. They appear repetitively in a region of our brains dubbed hippocampus, which is responsible for consolidating memories. Studying ripples is critical for us to better understand the underlying mechanisms of memory and many laboratories around the world focus their research on these events. It has been proved that interrupting ripples prevents memories from getting formed, and that prolonging them may improve memory.

The biological signal where we can see ripples is known as Local Field Potential or LFP. This is a kind of local EEG (or electroencephalogram). Neurons are constantly interchanging ions and this electric flow is translated into an electric potential that together with those emanating from surrounding regions, create an electrical signal which can be detected by our measuring devices. Therefore, the LFP is a signal that represents the overall neural activity of the many neurons that are around the area where the measurement electrode is placed. In the picture above, we can see 8 aligned electrodes recording the LFP at different equally spaced locations of the hippocampus.

Ripples are generated in the hippocampus, a very primitive brain region with roles in memory and spatial information processing (tasks such as creating a mental map of the world, orientation, etc). Hippocampal neurons are distributed in such a way that individual electric potentials add each other, making easier to see oscillations in the LFP.

Well! The goal of this hackathon is to develop an algorithm based on artificial intelligence and machine learning techniques, capable of detecting ripples in time series such as the LFP (sampled at 1250 Hz). In order to accomplish this, we will provide you a database with LFP recordings and the times of ripples manually annotated by our expert. You are free to use any supervised learning approach for addressing the challenge.

All these topics will be covered in detail in a series of lectures.

…the hackathon

Hackathon was a term originally coming from the hacker community and used to refer to a meeting of programmers enrolled in a collaborative challenge. These events usually last between two days and a week. Some of them have educative or social purposes, but in most cases the goal is to create functional software. In our case, the purpose is both to apply AI tools to neuroscience and also to teach these techniques in a didactic way :)

This term lumps together the concepts of marathon and hacker, as a reference to a collective experience with the common goal of developing applications in a collaborative way in a short amount of time. From an educational perspective, this kind of event represents a pedagogical tool, since it promotes collaborative work among peers aimed to problem solving; it focuses on the working process as an instance of learning and facilitates intrinsic motivation in participants.

Hackathons are, moreover, an increasingly common hiring method in business.

…what will I gain from this experience

• You will learn about hot topics, both in research and the business world.

• … and you will do so in a way that promotes hands-on practice, collaboration and fun.

• You will work with a multidisciplinary team in order to solved a problem in a short period of time.

• And, possibly, you will earn an award. Money can’t hurt you ;)

The BrainCode Games is funded by Sociedad Española de Neurociencia (SENC) and Universidad Autónoma de Madrid (UAM).

,