Publications

In this paper, we present a generic approach that can be used to infer how subjects make optimal decisions under uncertainty. This approach induces a distinction between a subject's perceptual model, which underlies the representation of a hidden "state of affairs" and a response model, which predicts the ensuing behavioural (or neurophysiological) responses to those inputs. We start with the premise that subjects continuously update a probabilistic representation of the causes of their sensory inputs to optimise their behaviour. In addition, subjects have preferences or goals that guide decisions about actions given the above uncertain representation of these hidden causes or state of affairs. From a Bayesian decision theoretic perspective, uncertain representations are so-called "posterior" beliefs, which are influenced by subjective "prior" beliefs. Preferences and goals are encoded through a "loss" (or "utility") function, which measures the cost incurred by making any admissible decision for any given (hidden) state of affair. By assuming that subjects make optimal decisions on the basis of updated (posterior) beliefs and utility (loss) functions, one can evaluate the likelihood of observed behaviour. Critically, this enables one to "observe the observer", i.e. identify (context- or subject-dependent) prior beliefs and utility-functions using psychophysical or neurophysiological measures. In this paper, we describe the main theoretical components of this meta-Bayesian approach (i.e. a Bayesian treatment of Bayesian decision theoretic predictions). In a companion paper ('Observing the observer (II): deciding when to decide'), we describe a concrete implementation of it and demonstrate its utility by applying it to simulated and real reaction time data from an associative learning task.

We have used event-related functional magnetic resonance imaging (fMRI) to characterize neural responses associated with emotional learning. Employing a classical conditioning paradigm in which faces were conditioned by pairing with an aversive tone (US), we compared responses evoked by conditioned (CS+) and nonconditioned (CS−) stimuli. Pairing 50% of the CS+ with the US enabled us to constrain our analysis to responses evoked by a CS+ not followed by a US. Differential evoked responses, related to conditioning, were found in the anterior cingulate and the anterior insula, regions with known involvement in emotional processing. Differential responses of the amygdalae were best characterized by a time by stimulus interaction indicating a rapid adaptation of CS+-specific responses in this region.

Abstract Biological self-organisation is a process of spontaneous pattern formation; namely the emergence of coherent and stable systemic configurations that distinguish themselves from their environment. This process can occur at various spatial scales: from the microscopic (giving rise to cells) to the macroscopic (the emergence of organisms). Self-organisation at each level is essential to account for the hierarchical organisation of living organisms (organelles within cells, within tissues, within organs, etc.). In this paper, we pursue the idea that Markov blankets – statistical boundaries separating states that are external to a system from its internal states – emerge at every possible level of the description of the (living) system. Through simulations, we show that the concept of a Markov blanket is fundamental in defining biological systems and underwrites the nature and form of interactions between successive levels of hierarchical structure. We demonstrate the validity of our argument using simulations, based on the normative principle of variational free energy minimisation. Specifically, we adopt a top-down approach to provide a proof of concept for the claim that the self-organisation of Markov blankets (and blankets of blankets) underwrites the self-evidencing, autopoietic behaviour of living systems.

How do humans come to acquire shared expectations about how they ought to behave in distinct normalised social settings? This paper offers a normative framework to answer this question. We introduce the computational construct of 'deontic value' – based on active inference and Markov decision processes – to formalise conceptions of social conformity and human decision-making. Deontic value is an attribute of choices, behaviours, or action sequences that inherit directly from deontic cues in our econiche (e.g., red traffic lights); namely, cues that denote an obligatory social rule. Crucially, the prosocial aspect of deontic value rests upon a particular form of circular causality: deontic cues exist in the environment in virtue of the environment being modified by repeated actions, while action itself is contingent upon the deontic value of environmental cues. We argue that this construction of deontic cues enables the epistemic (i.e., information-seeking) and pragmatic (i.e., goal- seeking) values of any behaviour to be 'cached' or 'outsourced' to the environment, where the environment effectively 'learns' about the behaviour of its denizens. We describe the process whereby this particular aspect of value enables learning of habitual behaviour over neurodevelopmental and transgenerational timescales.

ABSTRACT Hallucinations, including auditory verbal hallucinations (AVH), occur in both the healthy population and in psychotic conditions such as schizophrenia (often developing after a prodromal period). In addition, hallucinations can be in-context (they can be consistent with the environment, such as when one hallucinates the end of a sentence that has been repeated many times), or out-of-context (such as the bizarre hallucinations associated with schizophrenia). In previous work, we introduced a model of hallucinations as false (positive) inferences based on a (Markov decision process) formulation of active inference. In this work, we extend this model to include content – to disclose the computational mechanisms behind in- and out-of-context hallucinations. In active inference, sensory information is used to disambiguate alternative hypotheses about the causes of sensations. Sensory information is balanced against prior beliefs, and when this balance is tipped in the favor of prior beliefs, hallucinations can occur. We show that in-context hallucinations arise when (simulated) subjects cannot use sensory information to correct prior beliefs about hearing a voice, but beliefs about content (i.e. the sequential order of a sentence) remain accurate. When hallucinating subjects also have inaccurate beliefs about state transitions, out-of-context hallucinations occur; i.e. their hallucinated speech content is disordered. Furthermore, subjects with inaccurate beliefs about state transitions but an intact ability to use sensory information do not hallucinate and are reminiscent of prodromal patients. This work demonstrates the different computational mechanisms that may underlie the spectrum of hallucinatory experience – from the healthy population to psychotic states.

<h3>Background:</h3> The frequency of clinical isolation of non-tuberculous mycobacteria (NTM) in the Netherlands is increasing, but its clinical relevance is often uncertain. <h3>Objective:</h3> To assess the frequency and clinical relevance of isolation of NTM in four associated hospitals in a single region in the Netherlands. <h3>Methods:</h3> Medical files of all patients from whom NTM were isolated between January 1999 and January 2005 were reviewed retrospectively. Diagnostic criteria for non-tuberculous mycobacterial disease published by the American Thoracic Society (ATS) were used to determine clinical relevance. <h3>Results:</h3> 232 patients were found, from whom NTM were isolated from the respiratory tract in 91% of cases. Patients were mostly white men, with an average age of 60 years and pre-existing pulmonary disease. Fifty-three of 212 patients (25%) with pulmonary isolates met the ATS diagnostic criteria for pulmonary NTM disease; this percentage differed by species. Most patients were treated with rifampicin, ethambutol and clarithromycin. Treatment outcome for pulmonary NTM disease was suboptimal but differed by species: overall, improvement was seen in 67% of treated patients, but in only 50% of those with pulmonary <i>M avium</i> disease. Lymphadenitis was the most common extrapulmonary disease type. <h3>Conclusions:</h3> Twenty-five per cent of all patients with pulmonary NTM isolates met the ATS criteria. Clinical relevance differs by species. NTM isolation increases over time. Species distribution differs from that of neighbouring countries and the <i>M avium</i> complex isolates have traits different from those reported in the USA. Adherence to diagnostic and treatment guidelines can be improved.

Goal-directed actions are characterized by two main features: the content (i.e., the action goal) and the form, called vitality forms (VF) (i.e., how actions are executed). It is well established that both the action content and the capacity to understand the content of another’s action are mediated by a network formed by a set of parietal and frontal brain areas. In contrast, the neural bases of action forms (e.g., gentle or rude actions) have not been characterized. However, there are now studies showing that the observation and execution of actions endowed with VF activate, in addition to the parieto-frontal network, the dorso-central insula (DCI). In the present study, we established—using dynamic causal modeling (DCM)—the direction of information flow during observation and execution of actions endowed with gentle and rude VF in the human brain. Based on previous fMRI studies, the selected nodes for the DCM comprised the posterior superior temporal sulcus (pSTS), the inferior parietal lobule (IPL), the premotor cortex (PM), and the DCI. Bayesian model comparison showed that, during action observation, two streams arose from pSTS: one toward IPL, concerning the action goal, and one toward DCI, concerning the action vitality forms. During action execution, two streams arose from PM: one toward IPL, concerning the action goal and one toward DCI concerning action vitality forms. This last finding opens an interesting question concerning the possibility to elicit VF in two distinct ways: cognitively (from PM to DCI) and affectively (from DCI to PM).

The ubiquity and importance of narratives in human adaptation has been recognized by many scholars. Research has identified several functions of narratives that are conducive to individuals’ well-being and adaptation as well as to coordinated social practices and enculturation. In this paper, we characterize the social and cognitive functions of narratives in terms of the framework of active inference. Active inference depicts the fundamental tendency of living organisms to adapt by creating, updating, and maintaining inferences about their environment. We review the literature on the functions of narratives in identity, event segmentation, episodic memory, future projection, storytelling practices, and enculturation. We then re-cast these functions of narratives in terms of active inference, outlining a parsimonious model that can guide future developments in narrative theory, research, and clinical applications.