A problematic study of beliefs and dopamine

This 2015 Virginia Tech human study found:

“Dopamine fluctuations encode an integration of RPEs [reward prediction errors, the difference between actual and expected outcomes] with counterfactual prediction errors, the latter defined by how much better or worse the experienced outcome could have been.

How dopamine fluctuations combine the actual and counterfactual is unknown.”

From the study’s news coverage:

“The idea that “what could have been” is part of how people evaluate actual outcomes is not new. But no one expected that dopamine would be doing the job of combining this information in the human brain.”

Some caveats applied:

Measurements of dopamine were taken only from basal ganglia areas. These may not act the same as dopamine processes in other brain and nervous system areas.
The number of subjects was small (17), they all had Parkinson’s disease, and the experiment’s electrodes accompanied deep brain stimulation implantations.
Because there was no control group, findings of a study performed on a sample of people who all had dysfunctional brains and who were all being treated for neurodegenerative disease may not apply to a population of people who weren’t similarly afflicted.

The researchers didn’t provide evidence for the Significance section statement:

“The observed compositional encoding of “actual” and “possible” is consistent with how one should “feel” and may be one example of how the human brain translates computations over experience to embodied states of subjective feeling.”

The subjects weren’t asked for corroborating evidence about their feelings. Evidence for “embodied states of subjective feeling” wasn’t otherwise measured in studied brain areas. The primary argument for “embodied states of subjective feeling” was the second paragraph of the Discussion section where the researchers talked about their model and how they thought it incorporated what people should feel.

The study’s experimental evidence didn’t support the researchers’ assertion – allowed by the reviewer – that the study demonstrated something about “states of subjective feeling.” That the model inferred such “findings” along with the researchers’ statement that it “is consistent with how one should “feel” reminded me of a warning in The function of the dorsal ACC is to monitor pain in survival contexts:

“The more general message you should take away from this is that it’s probably a bad idea to infer any particular process on the basis of observed activity.”

The same researcher who hyped An agenda-driven study on beliefs, smoking and addiction that found nothing of substance was back again with statements such as:

“These precise, real-time measurements of dopamine-encoded events in the living human brain will help us understand the mechanisms of decision-making in health and disease.”

It’s likely that repeated hubris is one way researchers respond to their own history and feelings, such as their need to feel important as mentioned on my Welcome page.

The Parkinson’s patients were willing to become lab rats with extra electrodes that accompanied brain implantations to relieve their symptoms. Findings based on their playing a stock market game didn’t inform us about “mechanisms of decision-making in health and disease” in unafflicted humans. As one counter example, what evidence did the study provide that’s relevant to healthy humans’ decisions to remain healthy by taking actions to prevent disease?

The unwarranted extrapolations revealed a belief that the goal of research should be to explain human actions by explaining the actions of molecules. One problem caused by the preconceptions of this widespread belief is that it leads to study designs and models that omit relevant etiologic evidence embedded in each of the subjects’ historical experiences.

This belief may have factored into why the subjects weren’t asked about their feelings. Why didn’t the study’s design consider as relevant subject-provided evidence for feelings? Because the model already contrived explanations for feelings underlying the subjects’ actions.

http://www.pnas.org/content/113/1/200.full “Subsecond dopamine fluctuations in human striatum encode superposed error signals about actual and counterfactual reward”