**4. Theory of understanding: neuronal mechanisms of mental modeling**

The theory in a nutshell: Nervous system optimizes deployment of sensorymotor resources vis-à-vis varying external conditions, a part of the system that coordinates variations in the deployment of sensory motor-resources with variations in the conditions flow constitutes the first regulatory loop. Mechanisms of understanding operate on top of the first loop and optimize the organization of neuronal resources engaged in that loop, thus forming the second regulatory loop. Optimization in the second loop involves arranging neurons into coordinated structures manifested in coordinated mental models, as shown in **Figure 4**. Operations in the first loop are controlled by sensory-motor feedback while operations in the second one are decoupled from it. Feedback control makes resource deployment adaptive, self-controlled optimization in the second loop makes it self-adaptive [11]. First and second loops are stages of self-organization in the neuronal substrate. The first loop allows adaptation to compact sensory patterns extending over short time periods while the second one expands adaptation to dispersed patterns extending over indefinitely large time periods (prediction). limitations on the size of the neuronal pool and the amount of usable energy supplied per unit time drive the need to increase adaptation span while reducing energy costs, which boils down to a dual optimization criteria: minimize energy losses and the amount of energy consuming activities while maximizing prediction accuracy. Both criteria are subsumed under the notion of active inference [13, 14, 47].

This part will discuss the role of understanding capacity within the active inference framework, followed by detailed suggestions regarding neuronal mechanisms

**85**

theory, as follows.

*Brain Functional Architecture and Human Understanding*

**4.1 Active inference: from Aristotle to Friston**

*himself ignorant" (Aristotle, Metaphysics).*

causes leads to undesirable self-evaluation

that underlie the capacity and are responsible for the range of its operation, including the extremes. The part concludes by referencing experimental findings and ideas in the literature that might help in assessing biological plausibility of the

The opening line in Aristotle's *Metaphysics* states that "humans desire to understand" [48]. Lack of understanding engenders puzzlement, and failure to identify

*".. men of experience know that thing is so, but do not know why, while the others know the 'why' and the cause…. and man who is puzzled and wonders thinks* 

In a penetrating insight, Aristotle captures relations between experiences, surprise (puzzlement) and self-directed activities motivated by the desire to reach beyond the appearances (identify causes). Arguably, principles of active inference and variational free energy minimization advanced in [13] are congruent with those early insights. The principles assert that life in all its forms, from unicellular organisms to humans, is predicated on the organisms ability to use sensing to predict conditions in its environment and to conduct activities reducing the difference between the predicted and the actual experiences. Predictions require models of the environment, the variational free energy value determines, roughly, the (information-theoretic) distance between the current and the desired states that takes into account the difference between the predicted conditions and those that were actually sensed and the surprise experienced under the model (the smaller the

probability assigned by the model to the condition, the higher the surprise).

emergence of life, evolution, and brain operation into a seamless whole.

'Markov Blanket of node x' is a graph-theoretic term denoting a set of nodes in a directed graph that are connected to x by links incident to and from x. More loosely, the term can be used to denote a group of nodes in subnetwork X1 separating it from the rest of the network X. If links denote some form of interaction, Markov Blanket of X1 can be viewed as an interface through which internal nodes in X1 interact with their surrounds in X. On that view, Markov Blanket accords X1 a degree of (conditional) independence from X - a critical concept in the overall

The theory of life attributes emergence of life to spontaneous phase transitions in molecular networks ('primordial soup'), resulting in the formation of subnetworks that remain connected to their surrounds but acquire a degree statistical

Emphasis on activities directed at minimizing variational free energy underlies the notion of 'active inference,' which is best appreciated if contrasted to the idea of 'passive' inference expressed in Plato's allegory of the cave, as follows. Prisoners are chained to the floor inside a cave where they can see nothing of the outside world except shadows on the wall they are facing. The message is that people are caged inside their minds, senses are the only window into the world, and that window can be distorting. The allegory defines passive inference: prisoners can make guesses about the outside world but have no means to validate them or to use in any fashion. Active inference differs from passive inference in that it incorporates iterative actions on both the outside world and the model of that world that can lead to progressively improving guesses. Understanding involves a form of model manipulation that is best defined within the active inference theory through the notion of a Markov Blanket - the third conceptual pillar in the theory integrating ideas about

*DOI: http://dx.doi.org/10.5772/intechopen.95594*

present proposal.

*Connectivity and Functional Specialization in the Brain*

the basic catapult were reproduced in numerous designs.

response compositions to a few plausible alternatives.

the notion of active inference [13, 14, 47].

As observed by Jean Piaget [46].

*([46] p. 218).*

(body and the sensory-motor periphery, e.g. the tongue-ejecting mechanism) and regulatory component orchestrating activities within the body and at the periphery (i.e., animal's behavior in the environment). Both components undergo evolutionary development in the species while behavior regulation is amenable to adaptive changes in individuals during their lifetime (learning). In animals, learning is restricted to condition-driven variations within narrow envelopes of geneticallyfixed condition-response patterns and propensities. Condition-driven learning extrapolates from past precedents while mental modeling enables prediction and response construction under conditions having no such precedents. More precisely, models integrate past history within cross-coordinated structures so predictions produced by operations on the structure can be made consistent with (plausible under the entire past history) without repeating any of its elements. Moreover, models allow reproductive construction without replication, e.g., coordinations in

*"…mental coordinations succeed in combining all the multifarious data and successive data into an overall, simultaneous picture, which vastly multiplies their powers* 

Summarily, it has been suggested that a) the protohuman-to-human transition was associated with the emergent capacity to construct responses to dispersed stimuli patterns and b) the capacity is rooted in the mechanisms of mental modeling that represents such patterns as coordinated structures that suppress combinatorial explosion inherent in the construction process and reduce the number of

**4. Theory of understanding: neuronal mechanisms of mental modeling**

The theory in a nutshell: Nervous system optimizes deployment of sensorymotor resources vis-à-vis varying external conditions, a part of the system that coordinates variations in the deployment of sensory motor-resources with variations in the conditions flow constitutes the first regulatory loop. Mechanisms of understanding operate on top of the first loop and optimize the organization of neuronal resources engaged in that loop, thus forming the second regulatory loop. Optimization in the second loop involves arranging neurons into coordinated structures manifested in coordinated mental models, as shown in **Figure 4**. Operations in the first loop are controlled by sensory-motor feedback while operations in the second one are decoupled from it. Feedback control makes resource deployment adaptive, self-controlled optimization in the second loop makes it self-adaptive [11]. First and second loops are stages of self-organization in the neuronal substrate. The first loop allows adaptation to compact sensory patterns extending over short time periods while the second one expands adaptation to dispersed patterns extending over indefinitely large time periods (prediction). limitations on the size of the neuronal pool and the amount of usable energy supplied per unit time drive the need to increase adaptation span while reducing energy costs, which boils down to a dual optimization criteria: minimize energy losses and the amount of energy consuming activities while maximizing prediction accuracy. Both criteria are subsumed under

This part will discuss the role of understanding capacity within the active inference framework, followed by detailed suggestions regarding neuronal mechanisms

*of spatio-temporal extension, and of deducing possible developments"* 

**84**

that underlie the capacity and are responsible for the range of its operation, including the extremes. The part concludes by referencing experimental findings and ideas in the literature that might help in assessing biological plausibility of the present proposal.
