Small beginnings

Sometimes big things have small beginnings. Honestly I’m not even sure how to talk about it vaguely, but I’m going to give it my best shot. The short version is that I officially have my first dataset for an idea I had almost a year ago now, a “big idea” as it were. I’m confident that this will work, I haven’t had a chance to see what we got, but I did make a few initial checks to ensure everything was collecting correctly. Unfortunately, despite how monumental this feels for me, I can’t explain what we have… yet.

First, a quick intro since I’ve been slacking on regularly doing this. I’m a fifth year (as of this fall) PhD candidate in neuroengineering. I have my BS and MS in mechanical engineering and while they are both engineering fields, they are totally different skill sets. Last fall I took a full-time research position in a hospital where I had a fellowship to do a joint research project (this paper was the result). The last four years have been a journey and I’m not finished yet, but I feel like I’ve jumped ahead in my career a bit because I took that job, which brings me to the big news.

For a year now I’ve been hard at work on an idea I had. Hospital-PI, whos lab I research in, has been very supportive of the idea and has done his best to give me everything I need to make it happen. I promised him something amazing and I intend to deliver. So far we’ve used the idea I’ve had in a lot of different ways. None of them were the way I had originally intended. Part of the problem was funding, we need funding to pay for the time, effort, and equipment that goes into the project. But we’ve worked around it for the most part.

Thankfully, we’ve been testing the idea on other projects because it’s been refined and while it’s not the ideal case (the ideal case would cost money), we’ve managed to cobble something together that is impressive, but also better than where we had started. So when we finally got word that we found a participant that met a very specific set of criteria, criteria mind you that no one has met up until now, we jumped at the chance. Because we’ve had other false starts, I said I wasn’t going to get my hopes up until we had the data in hand and the experiment was done.

Well for those who follow along with my journey and my little 365 days of academia project, you will notice yesterday was yet another missed day. That’s because yesterday was a very busy day, not only did we collect the data, but I’ve already started to hammer away at the analysis. There’s a lot of work ahead for me, but there was a lot of worry going into this as well.

It’s only happened once so far, but we had a case where we plugged everything in and got absolutely nothing. Worse than nothing we had perfect line noise (as in electrical noise picked up by our equipment). There was nothing else, just line noise, which was disappointing. Since then I’ve been extra cautious with how we collect the data to make sure that doesn’t happen again. When we collected the data I had a bit of a panic because we didn’t see a few things I was expecting to find, mostly because everything is obscured by line noise so it just looks like a 60 Hz sine wave, scary to say the least!

But shortly afterwards I took a chunk of the data and looked at the power spectrum of it. Basically I was checking to see what the frequency content of the data looked like. If we had nothing, it would look like a flat line then a spike at 60 Hz, meaning once again we had line noise and line noise only. If we actually had something we would see a 1/F curve (F being frequency), basically a high(er) value at low frequencies that drops off as the frequency increases. We expect this when we record EEG because large groups of neurons have an easier time syncing up at lower frequencies causing higher amplitude signals at low frequencies and lower amplitudes at higher frequencies.

Well the good news is that we found that we had a 1/F curve with a huge spike at 60 Hz, which was expected. That means we have actual, for real, data that I can do stuff with now. It also marks the first dataset for this project and a possible chance to get funding for it now that we’ve shown it works. Or at least that will be the step AFTER I actually look through the data and figure out what we’ve got and what we can do with it.

There are so many implications to what I want to do with this idea, but if I talk about it that will give away too much information. So, for now at least, you’ll just have to take my word that if the data show something good then it will be a major advancement. I’m 93% confident that this will be one of the easier datasets I’ve had to work with, but we’ll see what we have once I start looking through the data.

Then I can finally celebrate properly.