Tag Archives: eeg

Brainwaves & deep learning

Book title: Neural Oscillations in Neural Networks

Subtitle: Top neural networks that work best with EEG data and top EEG task classifications and data signals that work best with neural networks

Although quite short (~100 paperback pages or ~600 Kindle pages), it took me about a year to finish this book. This volume consolidates the existing body of knowledge, along with my own knowledge and experience in the space. It is quite straight to the point and easy to read.

The content of this volume covers the fundamentals of EEG, neural networks, and lastly the combination of the two, more specifically which neural networks architectures (deep neural networks) that work best with EEG data, along with which EEG task classifications we can analyze in neural networks as well as which type of EEG data works as the ‘right’ input data into the various neural networks architectures.

This book is geared towards a few types of readers:

  1. If you are a student or someone who is curious about this space, this volume helps you save time learning the fundamentals of this space to help you make the right choices.
  2. If you are an expert in EEG but have no background in neural networks, this volume should give you enough basic fundamentals of neural networks, more specifically which deep neural networks architectures work best with EEG data.
  3. If you are an expert in neural networks but have no knowledge about EEG, this book should give you the fundamentals about EEG data, as well as where to find/get free EEG data, that are made available to the public via institutions or laboratories.

Below is the link to the Kindle and the paperback versions on Amazon and a quick intro video on my YouTube channels.

Amazon link: https://a.co/d/ey15Oss

Advertisement

EEG artifacts removal & ERPs in Emotiv EPOC & EEGLAB

If you are new doing research with neural oscillation (EEG) type studies, chances are you feel overwhelmed in this space, not only with the software/hardware products and solutions but also with the set up and the research design and methodology as a whole.

The good news is that there are forums and communities built around this space that are quite solid and strong. Yet learning from the forum discussions can be extremely time consuming and frustrating as well. This is one of the main reasons I wrote this book, to help researchers get started without needing to read volumes of documents spread across the many existing platforms.

While the book covers most of the essentials to help set up an EEG study, using Emotiv EPOC neuroheadset and EEGLAB, there are always some specific questions that might pop up. I try to answer those questions to best of my knowledge.

However, since there are many different nuances to methodologies, it is impossible to know the answer to all questions. In those cases, I highly recommend using your preferred search engines with the specifics because the engines tend to search the keywords in forums as well. If not, you can always go directly to forums and type your questions there.

In this blog post, I am including a couple of questions that tend to come up from time to time. I am also including a few helpful links at the end of this post.

Q1: Are there any educational materials to help guide in artifact removal when you were starting?

A1: I personally read lots of forum discussions (see Emotiv forums) or EEGLab forums. I also used Google search with specific questions as they came along during my data analysis.

In regards to artifacts removal, such as eye blinking, it can mostly be learned through practice. Artifact removal tends to be tricky and is mostly learned either in school labs or in the industry. If you have no experience in recognizing artifacts, I highly recommend reading the methodology sections of various published research papers and then start practicing until you feel comfortable in both recognizing as well as removing artifacts from your EEG bands data.

Q2: What are some of the helpful documentations with regards to EEGLab, specifically in importing events, setting epoch and saving the processed data (with removal of baseline and ran ICA).

A2: To my knowledge, depending on which EPOC series you have, the Emotiv Control Panel or the Emotiv Pro have features for ERP related tools. You can use these to set your ERPs. And when it comes to the baseline and ICA removals, you can do the same process as stated in my book.

Read more here: https://www.emotiv.com/glossary/evoked-response-potential

Also, Emotiv research page has links for research studies done at various universities. I used to scan through those papers’ methods chapters and sometimes found answers to specific questions as they came along. 


Also see:

https://forum.emotiv.com/
https://www.emotiv.com/category/independent-studies/
https://www.emotiv.com/category/publications/

Hope this helps. Happy holidays!