Which kind of philosophy should we follow in scientific studies?

Hi Mike and friends,

What is the proper logic of scientific research?

I am a beginner in neuroscience studies. When I read papers concerning emprical studies, I found that many studies adopted this kind of logic: raised a hypothesis and then to prove it. Most of time, that hypothesis would be supported by some experiment results. The authors would claim that their studies successfully proved something. And this secnario was just seen superfically from the description in the paper. We never know that hypothesis was indeed raised before the experiment or authors just adjusted the hypothesis to fit into the results later. Whichever the case, people seem to follow the same logic, namely to prove the hypothesis.

I am quite doubting this logic. The simple reason is that we could only have limited samples which were inadequate to prove something. According to the philosophy of Karl Popper, the reason of hypothesis to exist is to be rejected.So maybe the proper logic should be that we raised a hypothesis according to previous studies first and then tried to refute it using various approaches. However, it seems that very few studies adopted this logic.

I look forward to listening to your ideas.


Hi Jinwen. Good question – and an important one, because it gets at the nature of progress in science. Popper laid out an ideal scientific method, which everyone should strive for as much as possible. Unfortunately, not all studies do, and there are many weak or non-falsifiable hypotheses in neuroscience.

And of course, “proving” something is nearly impossible in a complex system like the brain, even with causal interventions.

I cannot speak for all, or even any single, paper. And I’m sure my empirical papers would suffer from the same weaknesses if you looked carefully enough. The fact is that neuroscience is very much an empirical science; it is not so difficult to develop weak hypotheses that provide a framework for designing experiments and interpreting results, but it is extremely difficult to develop really strong hypotheses that are falsifiable and that lead to a single mechanistic conclusion.

There are two ways to react to this: You can say that neuroscience is shit and a waste of time, or you can understand that neuroscience is a very young field, and new fields of science always fumble through until there is firmer ground upon which to build really good theories and hypotheses. Physics is often hailed as a model of theory, but physics is thousands of years old. Modern neuroscience as a serious scientific discipline is maybe 150 years old.

I lean towards being generous and adopting the second reaction. But of course, your comment does point to a serious weakness in neuroscience. That leads to a whole different discussion about the pressures and incentives in modern academic neuroscience research, which I think is at least partly to blame for people feeling the need to produce and publish preliminary results.

Anyway, great question and very thought-provoking; thanks for posting it :slight_smile:

Thanks Mike. Your points remind me that neuroscience is a relatively young discipline that cannot be looked as the same way as physics. I have to or should be pleased to embrace its imperfection in methods. This also means more opportunities.