PosterPDF Available

Hierarchical Bayesian models of social inference for probing persecutory delusional ideation

Authors:

Abstract

While persecutory delusions (PD) have been linked to fallacies of reasoning (1-4) and social inference (5-7), computational characterisations of delusional tendencies are rare. In this study, we examined individuals from the general population on opposite ends of the PD spectrum (Paranoia Checklist, PCL, 7) employing hierarchical Bayesian models of learning, in order to identify key mechanisms of aberrant social inference in persecutory delusions. Methods We pre-screened 1145 individuals from the general population, and included 151 participants who exhibited either low or high scores on the PCL across multiple time points. They made trial-wise predictions in a probabilistic lottery, guided by advice from a more informed agent and a non-social cue. Two experimental frames differentially emphasized causes of invalid advice: (i) the adviser’s possible intentions (dispositional frame), or (ii) the rules of the game (situational frame). We applied a set of computational models to participants’ trial-wise behaviour to examine possible reasons for group differences in behaviour. The model space consisted of three model families: (i) the hierarchical Gaussian filter (HGF; 8-9), (ii) a mean-reverting HGF, and (iii) the Rescorla-Wagner reinforcement learning model (10). Results Using Bayesian model selection, we found that the HGF explained participants’ responses better than other learning models across both groups and conditions. Participants used both social and non-social information sources when predicting the outcome (posterior probability of the winning model or p(r|y) = 0.92). Model parameters determining participants’ belief updates about the adviser’s fidelity and the contribution of prior beliefs about fidelity to trial-wise decisions, respectively, showed significant group-by-frame interactions: high PCL scorers held more rigid beliefs about the adviser’s fidelity across both experimental frames and relied less on social advice in situational frames than low scorers. Specifically, the parameter representing the tonic aspect of the log-volatility of the adviser’s fidelity and the response model parameter representing how much participants took advice into account (in contrast to non-social information) showed significant group-by-frame interaction effects (df=(1,150), F=5.05, p=0.02 and df=(1,150), F=6.45, p=0.01, respectively). Namely, high PCL scorers exhibited lower values of these parameters across both frames, whereas low PCL were significantly affected by the frame, showing reduced learning rates and an increased reliance on the advice in the situational compared to the dispositional frame. Discussion These results suggest that PD tendencies are associated with rigid beliefs about others’ intentions and prevent adaptive use of social information in “safe” contexts. This supports previous proposals of a link between PD and aberrant social inference.
RESULTS
M1 stats: p(r|y)= 0.5646, pxp=1
winning model | 3-level HGF
11
references
1 Freeman, Daniel. "Suspicious minds: the psychology of persecutory delusions." Clinical psychology review 27.4 (2007): 425-457.
2Stephan, Klaas Enno, and Christoph Mathys. "Computational approaches to psychiatry." Current opinion in neurobiology 25 (2014): 85-92.
3Friston, Karl. "A theory of cortical responses." Philosophical transactions of the Royal Society B: Biological sciences 360.1456 (2005): 815-836.
4 Rao, Rajesh PN, and Dana H. Ballard. "Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects." Nature neuroscience 2.1 (1999): 79.
5Lee, Tai Sing, and David Mumford. "Hierarchical Bayesian inference in the visual cortex." JOSA A 20.7 (2003): 1434-1448.
6Doya, Kenji. "Modulators of decision making." Nature neuroscience 11.4 (2008): 410.
7Fletcher, Paul C., and Chris D. Frith. "Perceiving is believing: a Bayesian approach to explaining the positive symptoms of schizophrenia." Nature Reviews Neuroscience 10.1 (2009): 48.
8Corlett, Phil R., et al. "Toward a neurobiology of delusions." Progress in neurobiology 92.3 (2010): 345-369.
9Stephan, Klaas E., Andreea O. Diaconescu, and Sandra Iglesias. "Bayesian inference, dysconnectivity and neuromodulation in schizophrenia." Brain 139.7 (2016): 1874-1876.
10 Stephan, Klaas Enno, et al. "Bayesian model selection for group studies." Neuroimage 46.4 (2009): 1004-1017.
11 Mathys, Christoph, et al. "A Bayesian foundation for individual learning under uncertainty." Frontiers in human neuroscience 5 (2011): 39.
12 Freeman, Daniel, et al. "Psychological investigation of the structure of paranoia in a non-clinical population." The British Journal of Psychiatry 186.5 (2005): 427-435.
13 Adams, Rick A., et al. "The computational anatomy of psychosis." Frontiers in psychiatry 4 (2013): 47.
analysis methods | model parameter comparison
Predictions &analyses:
I. Initial prior beliefs about adviser fidelity, !"
#$% &'()*+ ,-(./0
II. Tonic evolution rate at 2nd level (adviser fidelity), 1"&23'2 4567)8 45
III.Social weighting parameter,9&'()*+ :'()*+ ,-(./0
IV. Belief precision at 2nd level ,;"~'()*+ ,-(./0 ,+2.<0 (task phase
after volatility vs.before volatility, see figure 2).
2-way ANOVAs on MAP estimates for =>
?@$AB,C>,and D
3-way ANOVA on MAP estimates for E>
0
0.2
0.4
0.6
0.8
1
F
low PD, dispositional
low PD, situational
figure 3 | model comparison model space and winning model
***
**
*
* p<0.05, ** p<0.01, *** p<0.001
social weighting parameter
winning model | equations
Mapping of level 1 (
,G
, advice accuracy) to level 2 (
,H
, advice
validity):
I JKJ>L M J>NOG P M?J>BKQNOLR0(S)*773 ?JKT M J>B[1]
M J U K
KVWXY QN [2]
Mapping of level 2 (
,H
, advice validity) to level 3 (
,Z
, volatility of
adviser’s intentions):
I J>
?@B J>
?@QKB[ J\
?@B[ ][ C>L ^ J>
?@BT J>
?@QKB[0,+ ]J\
?@B : C>[3]
I J\
?@B J\
?@QKB[ C\L ^ J\
?@BT J\
@QK [0,+?C\B[4]
Belief update equation:
_=`
?@B ab
cdeO
f
cd
fg`QK
?@B [5]
The trial-wise updating of beliefs at any level of the hierarchy by
PEs depends on their weighting by sensory relative to prior
precision.Thus beliefs are updated more readily the more precise
the sensory information and the less precise (certain) the prior
belief. I.e. agents who are highly confident in their predictions are
less likely to update their model of the world in the face of
contradictory evidence.
analysis methods |model comparison
The model space
consisted of 6perceptual models paired with 3
response models [see figure 3] which proposed that participants’
beliefs are based on (i) both cue &advice information (integrated), (ii)
advice only (advice),or (ii) cue only, i.e. that only the given cue
probabilities (i.e., the pie chart) enter the belief-to-response mapping
(cue).
We performed random effects Bayesian model selection
(BMS) 10,
which treats the model as arandom variable in the population and
allows for estimating which proportion of the population is best
described by each of the models considered.
high PD, dispositional
high PD, situational
volatility
of
intentions
advice
validity
response model
advice
accuracy
{0, 1}
response
{0, 1}
-10
-8
-6
-4
-2
0
hH
initial beliefs about adviser fidelity
tonic evolution rate
DISCUSSION
Low PD participants
exhibited higher evolution rates (1")in the
dispositional compared to the situational frame, suggesting that they
updated their beliefs about the adviser’s fidelity more rapidly in the
condition when the adviser was emphasized as the potential source of
incorrect advice.
High PD participants
exhibited lower evolution rates (1")(than low
PD participants) which were similar across the two experimental
frames.This suggests that their belief-updating process was similar
across experimental frames, and consistent with the group-by-frame
interaction for the social weighting parameter Fthat they made less
use of social context when learning from advice.
High PD participants
also showed less variable belief precision (;")
estimates across experimental frames (vs.low PD) suggesting that
they might have amodel of the adviser’s intentions that is less
susceptible to contextual information (i.e., the frame).This is
corroborated by the reduced influence of experimental framing with
regard to how much high PD participants took into account social
advice (DB.
The results of the current study support the general idea that
delusions can be conceptualized as beliefs with overly high precision
13
and that belief precision might also be increased on the subclinical
level.
belief precision at 2nd level
E>~'()*+i,i-(./0i,i+2.<0i
0
0.2
0.4
0.6
0.8
1
!"
(
#
=
%
)
n.s. p > 0.24
experimental methods | procedure
summary:
We examined social inference with aprobabilistic advice-
taking task in asocial context and investigated the role of precision in
the belief updating process by manipulating (i) changes in the
association strength between advice and outcome (volatility) and (ii) the
social context by instructing participants about the task under one of
two experimental frames (experimentally-induced priors).
study design:B
etween-subject factorial design with 2groups (high PD
vs.low PD)and 2conditions (dispositional vs.situational frame).
sample:
We prescreened N=1’145 individuals from the general
population, and included 151 participants who exhibited either low or
high scores on the Paranoia Checklist (PCL) 12 across 3time points
spanning over >5 weeks. All cells (group xcondition) were matched
regarding age, gender, and education in years.
experiment:
Participants played an iterative probabilistic advice-taking
paradigm [see figure 2] under one of 2experimental frames,which
differentially emphasized causes of invalid advice:
(i) the adviser’s possible intentions (dispositional frame), or
(ii) the rules of the game (situational frame).
task:
Participants had to make binary decisions (blue vs.green) based
on asocial and anon-social cue (advice &pie chart) presented
simultaneously.Players could accumulate points with every correct
prediction.After predicting what colour would ‘win’ they were informed
regarding the real outcome.
Pie chart probabilities varied between 50:50,55:45,and 65:35.Advice
validity was varied across 210 trials [see figure 2].Correct predictions
result in the accumulation of points growing aprogress-bar toward the
targets displayed.Surpassing the targets earned participants an
additional smaller bonus (silver target) or large bonus (gold target)
additional to the standard reimbursement for their time.
Hierarchical Bayesian models of social inference for probing
persecutory delusional ideation
Andreea O. Diaconescu* 1, 2, Katharina V. Wellstein* 1, Lars Kasper 1, 3, Christoph Mathys 1, 4, Klaas Enno Stephan 1, 5, 6
1
Translational Neuromodeling Unit (TNU), Institute for Biomedical Engineering, University of Zurich & ETH Zurich, Switzerland,
2
Department of Psychiatry (UPK), University of Basel, Switzerland,
3
Institute for Biomedical Engineering, MR Technology Group, ETH Zurich & University of Zurich, Switzerland,
4
Scuola Internazionale Superiore di Studi Avanzati (SISSA), Trieste, Italy,
5
Wellcome
Centre for Human Neuroimaging, University College London, UK,
6
Max Planck Institute for Metabolism Research, Cologne, Germany
4j
4j
+(0k3lm3)S<
+(0k3lm3)S<
This generative model assumes that
top-down predictions about sensory
inputs are updated by experience via
prediction errors (PEs).which are
weighted with their inverse variance
(precision) 5,6 [see figure 1].
Based on this framework delusions
have been proposed to arise from
increased sensory precision of low-
level PEs, rendering these PEs
abnormally salient and leading to a
chronic surprise about sensory inputs.
It is assumed that the subsequent
formation of highly precise high-level
beliefs may represent acompensatory
response that is required to “explain
away” the low-level PE signals 7,8.
figure 1 | illustration of
predictive coding process 9
0
0.2
0.4
0.6
0.8
1
040 80 120 160 200
advice validity
trial number
task probability structure
inter-trial interval advice / cue decision outcome
3-4 sec 2 sec 5 sec 1 sec
figure 2 | social learning task and probability structure of advice validity
volatile task phase
CP Course code & behavior contact
analysis plan analysis
,3nG
,3
introduction | background
Persecutory delusions (PD)
are defined as an agent’s beliefs that others
are acting deliberately to cause them harm despite disconfirming
evidence 1.PD have been linked to fallacies of reasoning (e.g. jumping
to conclusions) and deficits in social inference.However, mechanistic
insight into the underlying processes are rare.
Generative modelling
has been proposed as apromising approach to
understand these underlying mechanisms, i.e. how sensory inputs are
probabilistically generated by hidden states of the world 2.
Bayesian brain theories and predictive coding provide such a
generative model and propose that the brain infers on the causes of its
sensations using ahierarchically structured model of the world 3,4.
,3:G
_o0730- &i+(0l3<3)S ,i
4j
R.p0<qir*70s
+?,tp[/Bi&+?pt,[/B+?,t/B
+(0l3<3)S
+(0l3<3)S
0
20
40
60
80
100
1st stable
phase
volatile phase 2nd stable
phase
0
20
40
60
80
100
1st stable
phase
volatile phase 2nd stable
phase
***
***
*
**
E>~'()*+i,i-(./0i**
... Computational approaches to neuropsychiatry such as those using a free energy principle framework have approached schizophrenic symptoms as rooted in a deficit of functional connectivity and hence of complex generative models (Montague, Dolan, Friston & Dayan, 2012), and have approached persecutory delusion (PD) as aberrant social perception related to impairment of generative models of others (Diaconescu, et al., 2019). These accounts may offer a satisfying account the of the failure of more realistic perception, and Friston, et al. (2016) have suggested that the relative persistence of false beliefs (delusion) in schizophrenia reflects an increased precision given to prior (false) beliefs in response to failures of attenuation of sensory information. ...
Preprint
Full-text available
Friston’s (2010) Free Energy Principle (FEP) offers an opportunity to rethink what is meant by the psychoanalytic concept of an object, or discrete mental representation (Ogden, 1992). The significance of such objects in psychoanalysis is that they may be superimposed on current experience so that perceptions are partly composed of projected fantasy, and partly of more realistic perception. From a free energy perspective, the psychoanalytic (person) object may be understood as a bounded set of prior beliefs about a ‘platonic’ sort of person, that provides a free energy minimising, evidence maximising, hypothesis to explain inference about – or dyadic interactions with – another. The degree to which realistic perception supervenes - relative to a platonic person object - will depend upon the precision assigned to the sensory evidence (concerning the person) relative to the prior beliefs about a platonic form. This not only provides a basis for explaining projection and transference phenomena, but also for conceptualizing a central assumption within object relations psychoanalysis. As an example, the paper examines the Kleinian theory of split good or bad part-objects as affectively-organized generative models (or platonic part-object models) formed in early infancy. This also provides a basis for building on work by Kernberg (1984, 1996) by conceptualizing the role of the part-object(s) in a continuum of reality testing, from mild errors in perception that are relatively easily corrected, through borderline affective instability and frequent shifts between part-object experience, to psychotic failures of reality testing, where Friston, Brown, Siemerkus & Stephan (2016) propose that aberrant precisions bias perception to high precision false beliefs (here cast as platonic part-objects), such as stable perceptions of others (and possibly oneself) as persecutory agents of some sort. The paper demonstrates the value that the history of clinical insights in psychoanalysis (including object relations) and a systems-based approach to the brain (including the free energy principle) can have for one another. This is offered as a demonstration of the potential value of an ‘Integrative Clinical Systems Psychology’ proposed by Tretter and Lӧffler-Stastka (2018) which has the potential to integrate the major theoretical frameworks in the field today.
... It seems possible that these genetic differences all contribute to limits on functional connectivity in various ways that in turn place limits on the potential complexity of generative models that may regulate affective functioning. Computational approaches to neuropsychiatry such as those using a free energy principle framework have approached schizophrenic symptoms as rooted in a deficit of functional connectivity and hence of complex generative models (Montague et al., 2012) and have approached persecutory delusion (PD) as aberrant social perception related to impairment of generative models of others (Diaconescu et al., 2019). These accounts may offer a satisfying account of the failure of more realistic perception, and Friston et al. (2016) suggested that the relative persistence of false beliefs (delusion) in schizophrenia reflects an increased precision given to prior (false) beliefs in response to failures of attenuation of sensory information. ...
Article
Full-text available
Friston's (2010) free energy principle (FEP) offers an opportunity to rethink what is meant by the psychoanalytic concept of an object or discrete mental representation (Ogden, 1992). The significance of such objects in psychoanalysis is that they may be superimposed on current experience so that perceptions are partly composed of projected fantasy and partly of more realistic perception. From a free energy perspective, the psychoanalytic (person) object may be understood as a bounded set of prior beliefs about a "platonic" sort of person that provides a free energy minimizing, evidence maximizing, hypothesis to explain inference about-or dyadic interactions with-another. The degree to which realistic perception supervenes-relative to a platonic person object-will depend upon the precision assigned to the sensory evidence (concerning the person) relative to the prior beliefs about a platonic form. This provides a basis for not only explaining projection and transference phenomena but also conceptualizing a central assumption within the object relations psychoanalysis. As an example, the paper examines the Kleinian theory of split good or bad part objects as affectively organized generative models (or platonic part-object models) formed in early infancy. This also provides a basis for building on work by Kernberg (1984, 1996) by conceptualizing the role of the part object(s) in a continuum of reality testing, from mild errors in perception that are relatively easily corrected, through borderline affective instability and frequent shifts between part-object experience, to psychotic failures of reality testing, where Friston et al. (2016) proposed that aberrant precisions bias perception to high precision false beliefs (here cast as platonic part objects), such as stable perceptions of others (and possibly oneself) as persecutory agents of some sort. The paper demonstrates the value that the history of clinical insights into psychoanalysis (including object relations) and a system-based approach to the brain (including the free energy principle) can have for one another. This is offered as a demonstration of the potential value of an "Integrative Clinical Systems Psychology" proposed by Tretter and Löffler-Stastka (2018), which has the potential to integrate the major theoretical frameworks in the field today.
Article
Full-text available
Previous research has stressed the importance of uncertainty for controlling the speed of learning, and how such control depends on the learner inferring the noise properties of the environment, especially volatility: the speed of change. However, learning rates are jointly determined by the comparison between volatility and a second factor, moment-to-moment stochasticity. Yet much previous research has focused on simplified cases corresponding to estimation of either factor alone. Here, we introduce a learning model, in which both factors are learned simultaneously from experience, and use the model to simulate human and animal data across many seemingly disparate neuroscientific and behavioral phenomena. By considering the full problem of joint estimation, we highlight a set of previously unappreciated issues, arising from the mutual interdependence of inference about volatility and stochasticity. This interdependence complicates and enriches the interpretation of previous results, such as pathological learning in individuals with anxiety and following amygdala damage.
ResearchGate has not been able to resolve any references for this publication.