Variants of uncertainty in decision-making and their neural correlates
Introduction
Interest in understanding the neurobiological mechanisms by which decisions are made has grown just recently. Such different disciplines like economics, psychology and neuroscience joint together in order to attempt a unified explanation of decision-making [23]. Yet, at that time, we are far away from providing a single general theory of human decision-making. But incorporating and bringing together the findings from the different disciplines concerning different aspects of decision-making will help in understanding the big picture. To keep in view with this aim, the present study deals with one aspect of decision-making, namely with the cerebral correlates of processes related to different variants of uncertainty.
In decision-making, it is rather important to anticipate consequences associated with different options or actions in order to decide favorably. However, there are events and circumstances in life that influence the outcome of decisions therewith making decisions uncertain. Uncertainty-causing events can be externally originating in the social and natural environment as well as internally originating within the individual. Kahneman and Tversky [30] discriminated variants of uncertainty subject to the perceived cause of uncertainty, i.e., externally attributed uncertainty and internally attributed uncertainty. The authors sub-divided the former into uncertainty based on frequencies and uncertainty based on propensities, the latter into uncertainty based on arguments and uncertainty based on introspective confidence, i.e., knowledge. Kahneman and Tversky [30] could show that the perceived cause of uncertainty is reflected in the way subjects try to resolve their uncertainty. A prominent coping strategy with uncertainty of frequency is to try to rate the probability of external events (e.g., “there is a 60% chance for rain tomorrow”). In contrast, a successful coping strategy with uncertainty of knowledge is an intensive memory search, most likely in combination with the attempt to get missing information from valid external sources (e.g., “I am quit sure that possums are mammals, but I do not know exactly”). That way, strategic processes are taken to reflect the respective variant of uncertainty. By using functional magnetic resonance imaging (fMRI), the brain may serve as an external criterion to test whether the phenomenological distinction drawn by Kahneman and Tversky [30] and several other researchers (e.g., [4], [28], [48]) is paralleled on the cerebral level.
Accordingly, by using fMRI, we aimed to investigate the neural correlates of processes related to externally attributed uncertainty in Experiment 1 (Exp. 1) and processes related to internally attributed uncertainty in Experiment 2 (Exp. 2) and beyond in as much the latter differ from those induced by externally attributed uncertainty. To reliably detect areas systematically varying with the level of the respective variant of uncertainty, we induced different degrees of uncertainty in the respective experiment, i.e., a parametric manipulation.
To date, there is a great number of studies investigating brain activations induced by uncertainty-related paradigms including rule induction and application [24], inductive reasoning [25], hypotheses testing [12], anticipation of monetary gains and losses [3], [32], response conflict (e.g., [50]), dynamical motion predictions [50] or guessing [13], [15]. Common to all these paradigms, is the prediction of uncertain events. The neural correlates reported with uncertainty are posterior fronto-median areas, including mesial Brodmann area (BA) 8, anterior mesial BA 6, corresponding to the pre-supplementary motor area (pre-SMA) and BA 32 often in company with BA 24, usually referred to as the dorsal anterior cingulate cortex. By contrasting rule learning with item learning, inductive reasoning with deductive reasoning, hypothesis testing with response selection, response conflict with no conflict or guessing with reporting, these studies suggest a general difference between processes under uncertainty and those which are quite certain. However, from where the uncertainty arises is neglected so far.
By using a parametric approach, we investigated whether processes related to the different variants of uncertainty are reflected by fronto-median activations (main effects) and if so, whether this brain activation also increases with increasing uncertainty (parametric effects). In a subsequent group comparison (between-subjects design), it was investigated whether networks underlying externally and internally attributed uncertainty differ significantly. Particularly, since storage and retrieval of acquired visuomotor associations are required for the suggested coping strategy in decisions under internally attributed uncertainty [30], fronto-parietal activations were expected in networks that sub-serve working memory functions [39], [17].
In addition to fronto-median areas orbito-frontal areas are known to be engaged in uncertain decisions, particularly those that are induced by reward expectancy [3], [9], [38], [43]. However, due to technical restrictions of the sequence in a 3T NMR system that usually causes signal voids [37], medial orbito-frontal activations could not be detected in the present experiments.
Section snippets
Participants
Sixteen (five females, mean age 24.9 years, range 21–35 years) right-handed, healthy volunteers participated in Exp. 1 and 12 volunteers (seven females, mean age 25.1, range 20–31 years) in Exp. 2. Informed consent was obtained from each participant according to the declaration of Helsinki. Experimental standards were approved by the local ethics committee of the University of Leipzig. Data were handled anonymously.
Stimuli, task and experimental session
Stimuli consisted of comic pictures showing UFOs differing either in color,
Behavioral results
Performance was measured by the rate of erroneously answered trials and reaction times of correctly answered trials (see Table 1). In Exp. 1, a repeated measures ANOVA with the five-level factor UNCERTAINTY yielded a significant main effect for both error rates (F(4,60) = 54.5; p < .0001) and reaction times (F(4,60) = 6.0; p < .001). A reduction of condition-independent uncertainty over the course of the experimental session was indicated by a significant decrease in reaction times (F(3,45) = 15.4; p <
Main task effects of externally and internally attributed uncertainty
When testing for the main task effect in Exp. 1 significant activations were elicited within the right posterior fronto-median cortex (pFMC), particularly within mesial BA 8/6, the right anterior insula, the cuneus, the cerebellar vermis extending laterally into the paramedian portion of the left cerebellar hemisphere and within a sub-cortical network, including the ventral striatum, the thalamus and the right midbrain area (see Fig. 2 and Table 2). Testing for the main task effect in Exp. 2
Discussion
Experiment 1 as well as Experiment 2 were designed to investigate whether processes related to different variants of uncertainty are reflected within the same brain areas. By using a parametric approach and therewith inducing different degrees of uncertainty, it was aimed to identify and compare the brain correlates of processes related to externally attributed uncertainty (Exp. 1), i.e., uncertainty of frequency, with those of processes related to internally attributed uncertainty, i.e.,
Acknowledgments
We thank Andrea Gast-Sandmann for the stimulus material, Jennifer Kittel for assistance with data analysis, Karsten Mueller Andre Szameitat and Stefan Zysset for support in fMRI statistics. This work was supported by the German Research Foundation (SPP 1107).
Note: this manuscript combines results from two papers which were recently published in Neuroimage [52], [53].
References (53)
- et al.
Functional imaging of neural responses to expectancy and experience of monetary gains and losses
Neuron
(2001) - et al.
Characterizing stimulus–response functions using nonlinear regressors in parametric fMRI experiments
Neuroimage
(1998) - et al.
Nonlinear regression in parametric activation studies
Neuroimage
(1996) - et al.
Dissociable contributions of pre-frontal and parietal cortices to response selection
Neuroimage
(2002) - et al.
Neural activity in the human brain relating to uncertainty and arousal during anticipation
Neuron
(2001) - et al.
Functional MRI studies of spatial and nonspatial working memory
Cogn. Brain Res.
(1998) - et al.
Activation of different anterior cingulate foci in association with hypothesis testing and response selection
Neuroimage
(1998) - et al.
Differential neural responses to positive and negative feedback in planning and guessing tasks
Neuropsychologia
(1997) - et al.
Ventromedial pre-frontal cortex mediates guessing
Neuropsychologia
(1999) - et al.
Analysis of fMRI time-series revisited
Neuroimage
(1995)
Event-related fMRI: characterizing differential responses
Neuroimage
Dissociable executive functions in the dynamic control of behavior: inhibition, error detection, and correction
Neuroimage
A midline dissociation between error-processing and response–conflict monitoring
Neuroimage
EEG activity during the performance of complex mental problems
Int. J. Psychophysiol.
Variants of uncertainty
Cognition
A region of mesial pre-frontal cortex tracks monetarily rewarding outcomes: characterization with rapid event-related fMRI
Neuroimage
Lipsia: a new software system for the evaluation of functional magnetic resonance images of the human brain
Comput. Med. Imaging Graph.
An investigation of the value of spin-echo-based fMRI using a Stroop color–word matching task and EPI at 3T
Neuroimage
The mid-ventrolateral pre-frontal cortex and active mnemonic retrieval
Neurobiol. Learn. Mem.
Right inferior pre-frontal cortex mediates response inhibition while mesial pre-frontal cortex is responsible for error detection
Neuroimage
The role of the anterior cingulate cortex in conflict processing: evidence from reverse Stroop interference
Neuroimage
A blueprint for target motion: fMRI reveals perceptual complexity to modulate a premotor-parietal network
Neuroimage
Subprocesses of performance monitoring: a dissociation of error processing and response competition revealed by event-related fMRI and ERP's
Neuroimage
Anterior cingulate cortex, conflict monitoring, and levels of processing
Neuroimage
Predicting events of varying probability: uncertainty investigated by fMRI
Neuroimage
Why am I unsure? Internal and external attributions of uncertainty dissociated by fMRI
Neuroimage
Cited by (160)
Unraveling Molecular and System Processes for Fear Memory
2022, NeuroscienceSpatial and chronic differences in neural activity in medicated and unmedicated schizophrenia patients
2022, NeuroImage: ClinicalNeural correlates of doubt in decision-making
2021, Psychiatry Research - NeuroimagingCitation Excerpt :In one such study relevant to doubt, Huetell et al. reported increased activation in the right inferior frontal gyrus (IFG), left middle frontal gyrus (MFG), bilateral anterior insula, right thalamus, right inferior parietal lobule (IPL), and the bilateral inferior parietal sulcus (IPS) in healthy adults during conditions of increasing uncertainty (Huettel et al., 2005). Volz et al. examined decision-making by differentiating between uncertainty that can be attributed externally such as the probability of the occurrence of an event and uncertainty that can be attributed internally such as internal knowledge about a winning rule (Volz et al., 2005). In this approach, external uncertainty is resolved by attempts at rating the probability of an event while internal uncertainty leads to a memory search to resolve uncertainty.
Aberrant anterior cingulate processing of anticipated threat as a mechanism for psychosis
2021, Psychiatry Research - Neuroimaging