Finding Neurological Evidence
Investigating affective neural population dynamics will lead to better classifica-tions of certain kinds of affective states. Experimentally observing which popu-lations specifically activate and modulate DSs, would significantly improve the neural understanding of affect. Using rTMS is a tested and innovative method to investigate these causal connections. The author suggests a larger scale be-havioural replication study to confirm the results from this study. In case of a successful replication, causal neuroscientific conclusions could be drawn through one or multiple of the following procedures.
Prior research has attributed the right dlPFC a key role in the moderation of higher order goals in the presence of distracting stimuli (Hare et al., 2011;
De Raedt et al., 2010; Korn & Bach, 2018). Exhaustion of the right dlPFC should hence alter affective states. Specifically, if rTMS is applied to exhaust the neurons in these dorsolateral areas temporarily, more risk avoiding DSs and less rnpmDSs will likely be used.
A potential between-subject experimental setup could look as follows: Neg-ative stimuli would be presented to the subjects on the trail before entering the forest. However, for some subjects, rTMS would be applied to deactivate the right dlPFC; for others only a placebo machine would be used. After a deacti-vation, the resulting affective state should lead to a lack of stimuli moderation and a strategy that leads to lower expected payoffs.
Next to the dorsolateral areas in the PFC, the dorsomedial areas have been attributed a key role. This area has been identified as the center of vital func-tions related to DSs, like computation of complex strategies through integration of many context-relevant variables (Korn & Bach, 2018). Certain types of de-pression have also been linked with inactivity in this region. In a long-term study, clinically depressed subjects could be treated with rTMS to increase the excitability of neural populations in the dmPFC. In a within subject design, the subjects would play the foraging game before and after a series of treatment ses-sions. A comparison of DS difference between their first and their second session could provide insights into how the dmPFC moderates DSs in the framework of affective states. Finding an appropriate subject pool may be difficult though.
The effect of long term rTMS treatment on healthy subjects has regrettably not received enough attention yet to make clear predictions about the impact of stimulation on DS use.
Next to creating causal connections through rTMS, future research also should aim to map affective state dependent DSs on a neural level. Evidence from this study combined with prior neuroscientific findings suggest that key affective regions like the vmPFC and the amygdala, dynamically interact with dorsal areas linked to strategy selection in the PFC. For example, increased cou-pling of the amygdala and dmPFC during ongoing anxiety have been observed (Vytal et al., 2014). Pairing this finding with the results of this study, it can be reasonably hypothesised that the interactions of these areas will correlate pos-itively with the variables for increased threat awareness, as well as negatively with the rnpmDS. An fMRI analysis developed around this hypothesis, could be in turn the basis for further studies observing more intricate affect dynamics and their relationship with DS variables.
Cross Disciplinary Models
This manuscript started with a discussion on concepts of affect and how they have developed over time. The verbalised intuitive experience of emotions is very different to the way economic utility functions describe valence oriented outcomes. In fact, it took a cross-disciplinary effort of economists and psy-chologists to develop the concept of decision value as a means to behaviourally formalise expected utility. Pioneering a young field, like neuroeconomics today, implies that the value of experimental results for future generations of scien-tists is ambiguous. Not all findings will reveal relevant information. Economists may argue that the level of detail neuroeconomics is modelling holds none of this long-term value. However, if one believes that the searching space for ex-planations is open ended, it is unwise to restrict one’s horizon. Models may seem complete at first, but are likely a level of abstraction with more detailed underlying functions. Psychological descriptions appeared to capture all major facets of behaviour, but neuroscientific measurements paint a new and more vibrant picture. Investigating the organ that is the main driver of our nervous system, took behavioural sciences one step further. The author believes that considering affective neuroscience in the context of computer science,
informa-tion theory or physics may add decisive parameters to current theories, to lift the understanding of the human condition to a new level.
For this study, computational DSs were defined and developed. These strate-gies were based on attributes associated with the stimuli presented. The proba-bility of success or danger are just individual evaluations of a stimulus without integration, whereas the rnpmDS optimises a monetary payoff over all informa-tion available. The human brain can clearly do better than simple evaluainforma-tions.
The rnpmDS on the other hand, is clearly out of reach to human cognition.
While the true nature of the computations used by humans is unknown, this ex-perimental study has demonstrated that affect is inherently linked to it. Certain aspects of positive affect lead to computations much closer approximating what the Markov decision process prescribes as optimal behaviour, pointing towards its intricate role in the integration of information.
No matter the approach, the evidence points to the importance of the dy-namic interactions between neural populations and incoming sensory informa-tion. This information causes ongoing changes to the affective state. The au-thor believes in the importance of these states for the higher cognition humans achieve with limited biochemical resources available. Mapping these processes through computational models needs to account for the affect state parameters.
Algorithms using affective parameters could elevate AI beyond current compu-tational models in real life scenarios. Events with large numbers of stimuli and more ambiguous time horizons than this experiment, likely require the guidance of inherent dynamic computational states. These are precisely the kind of states the human body seems to adapt, maintain and utilise so effectively.
A more practical and abstract way of looking at affective DS interactions could be through an information theoretical approach. Information theory con-ceptualises the storage and transfer of structured data (Shannon, 1948). In-formation theory is a powerful tool for neuroscientists because of some distinct features. Prominently the theory does not make any assumptions about the structure of the underlying data. In other words, information theory uses a model-free approach, not depending on any underlying functions. This is specifi-cally helpful in a field like neuroscience, where the underlying functions are often not well understood (Timme & Lapish, 2018).
Information theory hence provides an alternative way to describe neural activity. It is a toolbox to map how stimuli get captured and processed. The brain retains outside information, combines it with existing information and stores it as short and long-term memory. Sensory information, if old or new, is converted into behavioural information. Evidence from the foraging game points to the way affect interacts with these information flows. The internal state interacts with the incoming sensory information to select which information enters the cycle of memory and behaviour. Mapping how information enters the brain and where it travels could provide insightful mathematical models of neural activity in the future.
Lastly, this study seeks to discuss interdisciplinary efforts between physics, specifically the rules of quantum mechanics, and neurobiology. Unifying the two concepts on the basis of current scientific knowledge would be a truly difficult
task. Explaining the complex and little understood field of neuroscience with the strangeness of quantum mechanics may seem to not hold much merit. Rather are neurobiological phenomena usually described on a level of abstraction that is measurable and comprehensible. This is a fruitful approach, if a researcher wants to prevent drifting into meta-science and speculation.
Nevertheless, the principles of quantum mechanics explains the microscale of the universe, including the particles that make up the nervous system. Hence, it is crucial for any natural science to not disregard the framework it is operating in. Talking about basic emotions or a contrast of cognition and affect, does not make sense in the roam of quantum systems with their inherent randomness.
Current affective models are theoretical concepts, which should guide research as long as they provide utility. If, however, they are restricting thinking patterns and hence progress, they should be discarded with the notion that they are modelling on a level of abstraction, which is unsuitable for describing newly discovered mechanisms and phenomena (Koch & Hepp, 2006).