The Role of Neuroscience in a Functional Analysis and Understanding of Behaviour

A repurposed essay and an example of analysis in Psychology and Neuroscience

Do we need neuroscience to attain a functional understanding of human behaviour? By imagining a generic case study, and examining a biological and psychosocial approach to analysing behaviour, I’ll argue that neuroscience is necessary but insufficient.

Originally published on Substack. Click this link to follow me there.

‘Functional understanding’ is defined broadly as an understanding of our behaviours and motivations for them. That definition is analogous to a Computational level of analysis — one of three levels of analysis concisely described by Churchland and Sejnowski, and first outlined by Marr and Poggio. These analyses assume the mind can be thought of as an information processing system, and that the first step is to understand what problem system is solving and why.

A Computational level of analysis decomposes tasks into their most important steps. The Computational level is followed by an analysis at the level of algorithms and then physical implementation. After determining the steps required to complete the task, the system determines which algorithm is required to process information and execute an output once a decision has been made about the inputs. Finally, the system implements the algorithm.

As an analogy, we might consider a musician at various levels of analysis.

  1. Computation: defining the problem and goal, and success and fail states for that goal. For a musician, this step can be as simple as reading the notes on the page.
  2. Algorithm: specifying a procedure to achieve the goal and the expected outcomes of certain actions. For the musician, this means deconstructing the notes into a series of tasks to be performed and determining which actions lead to the right outcomes. The musician forms a series of smaller sub-routines: playing the notes, over a certain period of time, in a certain order, etc.
  3. Implementation: the musician runs the algorithm with their chosen instrument, successfully acting out the subroutines.
  4. Implementation is multiply realisable, the musician could use their voice, a guitar, a flute, or any other sound producing equipment. The chosen method of implementation restricts the algorithmic possibilities and, therefore, the range of likely computation, but it doesn’t fully determine specific actions.

If a musician can only use a bass-guitar, then they are limited to algorithms which work on bass-guitars and bass-guitars alone. But, creative freedom remains at the computational and algorithmic levels: the speed of play, the grouping of notes, pitch, etc.

These levels are also outlined and put to use in Part II of ‘Individuals and Groups: Exploring the right level of analysis’.

Reasoning by Fictitious Case Study

Otto, 17, is an only child living at home with his mother. His mother is unwell and has not been able to work for a few years. Otto has a mild learning impairment but attends sixth form and works part-time. Otto’s friends from high school do not attend any further education programmes. Two friends work full time and others found alternative methods of earning. Otto was a recreational drug user but now has a substance use disorder. He continues to use cocaine even though he experiences serious, negative consequences for doing so (American Psychiatric Association, 2019). Otto is addicted to cocaine. Most of Otto’s friends are recreational users or addicts, too.

There are two primary approaches to understanding Otto’s addiction: a biological (neuroscientific) approach and a psychosocial approach.

Finding a neuroscientific explanation for Otto’s behaviour means describing what causal mechanisms in Otto’s brain are motivating his behaviours and determining how he responds to stimuli. Precisely: what about Otto’s brain can help us understand why Otto desires cocaine and takes the necessary steps to use cocaine?

What follows is an example of analysis and not a definitive statement of facts.

Pascoli et al. claim that stimulating ventral-tegmentum area dopamine (VTA DA) neurons in mice “induces behavioural and cellular hallmarks of addiction, indicating sufficiency for the induction and progression of the disease.” (Pascoli, Terrier, Hiver, & Lüscher, 2015, p. 1054) To provide evidence for their claim, Pascoli et al. report their findings from a variety of optogenetic tests focussing on dopamine neuron activation. They found that 20% of mice were “resistant” rather than “sensitive” to punishment and will continue to self-stimulate VTA DA neurons despite negative consequences. Furthermore, Pascoli et al. claim to have found neural correlates for this addictive behaviour. They found that neural activity in the Orbitofrontal Cortex (OFC) is best associated with resistance to punishment in resistant mice. This finding was supported by further chemogenic tests using Clozapine-N-Oxide to inhibit activity in the OFC. Of the resistant mice, 5 out of 16 were resistant when OFC activity was inhibited and 14 of 16 mice were resistant when OFC activity was normal.

In summary: Pascoli et al. found evidence in favour of a purely neuroscientific explanation of addiction. Opportunities for VTA DA self-stimulation, paired with individual differences in OFC excitability, leads to repeated self-stimulation which is resistant to negative consequences.

For Otto, then, there is at least one possible explanation for his behaviour. He has been presented with an opportunity for VTA DA self-stimulation (access to cocaine) and he is unfortunate enough to also have enhanced OFC excitability, so he is resistant to changing his use despite negative consequences in his life.

A psychosocial approach may construe behaviour as a function of the person and their environment: B = ƒ(P,E) — also known as Lewin’s equation. Finding a psychosocial explanation for Otto’s behaviour means describing his motivations and methods of cocaine use as a function of him and his wider environment.

Lewin’s equation offers two factors which can be manipulated: the person and their environment. The person is divisible into two further factors, psychology and biology. Biology influences psychology, for every psychological event there are likely to be contiguous and contingent biological processes. The biological approach has already provided one such possible example for Otto: coinciding with his choices to use cocaine were underlying biological processes. However, while these biological processes may be necessary for his addiction, the psychosocial approach denies that they are sufficient alone.

The two, primary causes of insufficiency are that biological approaches fail to account for the environment and typically rely on animal models, which cannot account for the complexity of human psychology and decision making. Particularly, Pascoli et al. have used animal studies with low ecological validity (even among animal studies).

I’lll tackle the problem of ecological validity first. Simply put, unlike Pascoli et al.’s mice, Otto doesn’t have access to an unlimited supply of an unusually effective method of dopamine stimulation. Furthermore, even if Otto did have access to the same or equivalent methods of stimulation, the results are only interpretable all else being equal. In a complex social environment, little remains the same — even for a single individual over time. Otto’s life is subject to all manner of potential confounding variables. He may have social influences to use cocaine, economic pressures contributing to his desire to use cocaine, an environment that makes cocaine readily available, no adequate healthcare systems to help him, or any manner of other factors.

Beyond ecological factors, Otto has a more complex psychology than mouse models will capture. In our example, Otto has a mother who is sick and unable to work, which presents an enormous potential source of stress for which cocaine use may be a coping strategy. He may also be using cocaine as a means to be more productive while he juggles work and study. Perhaps Otto uses cocaine as a means to socialise and keep partying. The table below, produced by Boys et al., offers a number of self-reported reasons why people use substances, including cocaine.

In the absence of these psychosocial motivations, one could question whether Otto would have ever turned to cocaine, or whether he would be unable to appropriately regulate his use. Certainly, if Otto didn’t have access to cocaine, then he wouldn’t be addicted to it, irrespective of any biological predispositions. Perhaps, that’s true of each of his potential motivations. If it is, then coming to a functional understanding of Otto’s behaviour (an understanding of what problem Otto is trying to solve and how cocaine solves it for him) is superior.

Given Otto’s addiction, Otto will also be trying to solve the biological problems addiction causes. Furthermore, as Otto uses drugs, he may experience alterations in gene expression and changes in neural plasticity which facilitate his continued use. But beyond that, a biological account of Otto’s drug use fails provide any further insight into what problems Otto was originally trying to solve. There must be a reason for the first time Otto used cocaine. Otto’s implemented solution to his problem was multiply realisable, he could have implemented the algorithm differently or deconstructed the problem into a different set of tasks to be implemented. Just as in the previous illustration of the three levels of analysis, the musician could have made other choices and so could Otto.

Neuroscience can inform us of the limitations on our possible range of actions, but as of yet, it cannot fully explain the motivations for our actions. I agree with Churchland and Sejnowski, “neurobiological data provide essential constraints on computational theories…” (Churchland & Sejnowski, 1988, p. 242). But those constraints don’t prescribe only one course of action. While we act with consideration to what is possible — what steps we can physically implement to solve our problems — what is possible only partially determines what we do.


American Psychiatric Association. (2019). Addiction and Substance Use

Disorders: What is Addiction? Retrieved from American Psychiatric Association:

Boys, A., Marsden, J., & Strang, J. (2001). Understanding reasons for drug use amongst young people: a functional perspective. Health Education Research, 16(4), 457- 469.

Churchland, P. S., & Sejnowski, T. J. (1988). Perspectives on Cognitive Neuroscience. Science, 741–745.

Lewin, K. (1937). Principles of Topological Psychology.

Marr, D., & Poggio, T. (1977). A Theory of Human Stereo Vision. MIT Artificial Intelligence Laboratory.

Nestler, E. J. (2013). Cellular Basis of Memory for Addiction. Dialogues in Clinical Neuroscience, 431–443.

Pascoli, V., Terrier, J., Hiver, A., & Lüscher, C. (2015). Sufficiency of Mesolimbic Dopamine Neuron Stimulation for the Progression to Addiction. Neuron, 1054–1066.