Neuroscientific measurements and expensive, probing neuroscientific experiments have dominated research methods in psychology and psychiatry in recent decades. A new position paper from Yael Niv, a cognitive neuroscientist at Princeton University, soberly evaluates the outcomes of this shift away from studying behavior and whole persons. Niv urges fellow scientists, journal editors, and funding bodies to return to studying behavioral research to better understand brain function and get back to fulfilling the promise of improved patient care.
With fervor commentators have called “religious,” belief in always forthcoming breakthroughs in the neuroscientific understanding of the brain has channeled extensive funding and effort into its experimental methods. Even as many have suggested that such belief is philosophically unfounded and fails to explain psychopathology, trends in publicly-funded research reveal that these methods continue to be prioritized. Neural data collection is emboldened by projects like the Brain Initiative and Human Connectome Project, and behavioral, humanistic, and psychoanalytic research in psychology has been confined to the filing cabinets of unscientific or irrelevant methods.
The demotion of behavior research is not only reflected in prominent neuroscientific journals’ decisions to reject behavioral studies out of hand. It also shows up in discourses of “neural agency” by researchers who describe decisions made by neurons, not by people. News coverage of neural imaging and measurement studies make neural attributions common among laypeople who regularly attribute their actions to the actions of their brains.
This is why Princeton computational cognitive neuroscientist and feminist Yael Niv’s pre-printed article, “The primacy of behavioral research for understanding the brain,” is so significant. She suggests that the neural approach is reductionistic and fundamentally flawed, exploring how we got here and why well-crafted behavioral paradigms offer superior insights into the brain.
She begins by describing behavioral studies that illuminate brain processes far beyond what neural research has achieved. In the pertinent example of a rat trained to run from the base of a T-shaped maze to its right arm where food is located, no amount of recorded brain activity could answer a rather simple question: Is the rat’s strategy based on egocentric cues (i.e., turning relative to its own body) or external cues (i.e., turning based on what is afforded in the room, like light from a window)?
“By manipulating the brain, we can find out what brain areas can affect behavior on this task, but not what brain areas do affect behavior as the rat is making its right turn,” she writes.
In essence, perturbing an area in the brain teaches researchers how to refine their perturbation and manipulation techniques, but not about what is going on in the absence of experimental effects (which is to say, reality). To answer this strategy question, what was required instead was to turn the maze around such that it pointed north rather than south.
This simple technique indicated the use of external, peripheral cues and led to an understanding of comparative training lengths for behavioral change. Only later were computational models constructed to explore how the rats may learn to transition from one strategy to another.
This example highlights the primacy of behavioral findings related to learning (i.e., that it occurs through error-correction) to which neural computational models may later be applied to refine explanation and prediction with greater precision.
“Arguably, the most interesting findings from neural recordings or perturbations are when those are coupled with incisive, hypothesis-driven behavioral experimental designs. It seems that neuroscience alone is not nearly as revealing as are a combination of a telling behavioral design, neural recordings, or perturbation of neural function (e.g., through lesion, inactivation, or stimulation), and a computational model that states the hypotheses to be tested precisely.”
Niv claims, against the assumptions of funding bodies and scientific journals, that behavioral data is not only as important as neural data but that it is necessary for neuroscientific data to glean insights about the brain and to answer the questions neuroscientists tend to focus on. While it is obviously the case that studying the brain is useful for neuroscience itself, Niv suggests that neuroscientific measurements play only a confirmatory role in understanding cognition.
She suggests that reversing the current research pyramid (in which neural measurement is primary) is necessary to avoid continuing on the expensive and futile path of using neuroimaging studies to understand complex cognitive functions. Niv adds:
“Assuming that a process such as decision making can be understood by looking at single neurons, or even their ensembles is like attempting to understand why people in Australia drive on the left side of the road from an examination of their DNA. Neural firing patterns are the wrong level for investigating many pressing questions in neuroscience.”
Two illusions may have contributed to the reverse prioritization of neural and behavioral studies in recent years. First is the view that neuroscientific data is more “objective” than using a computational theory to interpret behavioral data. Another misperception is that behavior is already well understood or “solved,” and thus no longer interesting. On this assumption, psychologists sought to dive more deeply into the brain, which they believed to be the fundamental generator of behavior. Perhaps this was premature. Niv urges those who want to know more about behavior to study behavior rather than a decontextualized brain.
Niv sees the economic “sunk cost fallacy” at work in the continued funding of neuroscience over psychological research. Behavioral work is cheaper and easier than neuroscientific work, which leads many to consider the latter more valuable and rewarding.
Also, Niv points to methodological problems for invasive techniques deemed more powerful, like attempting causal manipulations by silencing or activating a set of neurons, and draws attention to the forgotten “super-power” of psychologists. Namely, the ability to design ingenious behavioral experiments that isolate processes and manipulate them at the less invasive level of the whole behaving organism.
Ultimately, Niv admonishes neuroscientists to work more cheaply and effectively in the name of taxpayer money, mental health patients, and increased opportunities for students and faculty alike. She powerfully advocates accelerating the process of understanding the mind by reasserting behavioral research as the undisputed basis of cognitive neuroscience.
Niv, Y. (2020, October 22). The primacy of behavioral research for understanding the brain. https://doi.org/10.31234/osf.io/y8mxe (Link)