Worth applying MD/PhD this cycle?

[SpanishMusical]

Hey everyone,

After a lot of deliberation, I've decided that I want to apply MD/PhD. I was hoping someone could give me some guidance as to if I should apply this cycle or not, and any schools they would recommend. My main interests are in computational/big data research, with a focus in disease biomarkers and nosology/disease classification (especially psychiatric diseases, although I could definitely get interested in this type of research in other fields.

My stats:
GPA: 4.00 (Spanish BA, Neuroscience BS)
MCAT: Cancelled (for now) due to COVID, but on AAMC practice tests I've been around 523
State of residence: IN
Race/gender: White and Hispanic (of Spanish descent), male; I do not consider myself URM
Undergraduate: State school, currently a Junior

Research:
- 3.5 years in my current lab (cognitive and clinical neuroscience); about 10 hrs a week, with work for 2 summers
- One third-author publication in a journal with IF ~5
- Working on an honors thesis to use electrophysiological data and machine learning to determine if distinct biotypes exist within psychosis diagnoses
- Paid EEG technician for one semester (until the grant ran out)
- Played an integral role in organizing over decades' worth of data into a database as well as writing code for easier data extraction and use

Clinical experience:
- English-Spanish medical interpreter at a local free clinic
- 53.5 hours patient intake at this clinic (after becoming EMT certified)
*I had to switch activities after the free clinic underwent restructuring and could no longer take on volunteers
- Hospice volunteer (around 10 hours; was cancelled early due to COVID)
- Shadowing: (16 hours for US MD only, 31 hours total for US and non-US MD, 65 hours total in healthcare)
- 8 hours ophthalmologist
- 8 hours ENT, 8 hours ER (technically shadowing a nurse)
- ~8 hours during my interpreting at the free clinic (again, technically shadowing nurses)
- 12 hours on an ambulance
- 15 hours in orthopedics in Spain (not sure if relevant, as it's not in the US); was hoping to get a little more before applying, but most was cancelled due to COVID

Volunteering/other:
- Weekend Spanish language school: 43 hours total
- 123 hours at front desk of free clinic
- Classical and Flamenco guitar for fun

I worry that my research, from a purely numbers perspective, is lacking. I've only worked in one lab; that said, I understand my projects well, including independently coming up with my honors thesis, and feel that I will have a very strong letter of recommendation from my PI.

Any thoughts on if I have a shot at MD/PhD this cycle, and if I do, at what schools?

---

Members don't see this ad.

 

[deleted728364]

Definitely looks like you’ll get a few acceptances, whether they’ll be at top programs that your stats allow is a question for the adcom members on here.


But a q about ML for EEG - is 90%+ performance despite the SNR? EEG is insanely noisy but I haven’t seen any popular approaches to denoising, but that would be cool in order to use stuff like deep learning.

 

[SpanishMusical]

Definitely looks like you’ll get a few acceptances, whether they’ll be at top programs that your stats allow is a question for the adcom members on here.


But a q about ML for EEG - do you find that there’s a high enough SNR for 90%+ performance? EEG is insanely noisy but I haven’t seen any popular approaches to denoising, but that would be cool in order to use stuff like deep learning.

Thanks for the response! Tbh if I can get a strong background in research while still being able to see patients at the end of everything, I'll be happy, so pedigree doesn't matter too much to me.

And that's an interesting question for sure, and one that I may have oversimplified in my original post. If one were to use the raw EEG data, yeah, I think you'll be really hard-pressed to get anything useful out of the raw data alone. At least in my experiment (and in a good number of others I've seen), we won't be using raw EEG data so much as various time-frequency parameters from subjects performing a certain task (namely, oscillatory power and phase synchrony in various frequency bands). We can then use these parameters to create clusters of people (which is the "machine learning" I'm referring to here, although this is admittedly a pretty broad definition) and see if the groupings differ on certain self-report measures that should theoretically be different based on which parameters end up driving the clusters (and in which frequencies). This would be super cool to use neural nets on, though, especially if someone had treatment response data with all the other parameters (I can dream!).

 

[deleted728364]

I have virtually the same setup in my lab. I do think that using summary stats leaves out a lot of useful data, but i guess eeg is just too noisy for anything else.

 

[Lucca]

if ur MCAT is near ur practice tests you should be able to get MD/PhD IIs. If you want to do the dual degree, I'd go ahead and apply this cycle. With COVID the next year will not be very productive for anyone and will delay training yet another year; not worth it given you're already pretty competitive as is. I think you will actually be more competitive for MD/PhD programs than MD only as your ECs other than research will not be on the competitive end for T20 and up even though your stats definitely are.

It's fine to only work in one lab as long as you can demonstrate command over your project and have a strong LOR from your PI. Could your research experience be stronger? Yes, it could be improved but the chances of doing that effectively over the next year with COVID are low and investing 1-2 years pre-PhD to get into a better MSTP isn't worth it IMO unless you're finishing college ridiculously early (and even then...). Better to get in as quickly as possible once the app is competitive and then save those 1-2 years to invest later in training when it really counts.