An Active Inference model on implicit awareness in Alzheimer’s disease
Poster No:
1081
Submission Type:
Abstract Submission
Authors:
RICCARDO MARAMOTTI1, Thomas Parr2, Daniela Ballotta1, Chiara Carbone1, Najara Iacovino1, Manuela Tondelli1, Giuseppe Pagnoni1, Giovanna Zamboni1
Institutions:
1University of Modena and Reggio Emilia, Modena, Italy, 2University of Oxford, Oxford, United Kingdom
First Author:
Co-Author(s):
Introduction:
Anosognosia (i.e. lack of illness awareness) is a common feature of neurodegenerative diseases, marked by a failure to recognize one's cognitive, emotional, and behavioral impairments. Recent findings suggest the presence of an "implicit" knowledge that operates at an automatic, preconscious level and may influence behavior even in patients who lack explicit awareness of their deficits (Mograbi & Morris, 2013). This implicit component can be examined using the Emotional Stroop experiment, a color-naming task that measures interference from emotionally charged words (Martyr et al., 2011). Our group (Tondelli et al., 2022) has shown that patients with Alzheimer's Disease (AD) without explicit awareness exhibit a differential activation for dementia-related words compared to emotionally negative words in the Posterior Cingulate Cortex – a key node of the Default Mode Network. This pattern was not featured in AD patients with explicit awareness. However, rection times showed large variability in AD patients and posed a challenge for the use of conventional statistical methods in assessing group differences. We adopted therefore a more sophisticated (and mechanistic) approach, by building a generative model of response times based on Active Inference. Here, we report the results of the predictive simulations necessary to assess the validity of the model's assumptions against the features of the collected data.
Methods:
A computerized Emotional Stroop task was administered to 37 healthy volunteers (age 65.4 ± 6.6) and 40 AD patients (age 72.3 ± 6.7). Participants were presented with colored neutral, negative, and disease-related words, and instructed to press one of three buttons corresponding to the color of the word (red, green, or blue). The experiment included 216 randomized trials. Each word was displayed until a response was issued (up to a maximum time of 1400 ms), with a 600 ms intertrial interval. A single-level Partially Observable Markov Decision Process (POMDP) model grounded in the Active Inference framework was developed to simulate behavior. Active Inference integrates perception, action, and decision-making treating behavior as an inferential process where actions are probabilistically chosen based on prior knowledge and sensory input (Parr et al., 2022). Our model included four hidden state factors: word meaning, word color, mental action (i.e. the modality chosen to respond with) and task sequence (with three levels, before the word appears, while viewing the word and while responding). For more details, see Figure 1. Four subject-specific parameters were included: one for learning effects, a temperature term for random variability, and two for the salience of negative and disease-related words.
Results:
As expected, the patients' actual reaction times were non-normally distributed, with a large variance. The simulated data were similarly distributed and exhibited the well-known speed-accuracy trade-off and the expected differences across the experimental conditions. A specific slowing of reaction times for disease-related words compared to negative words, which we assume indicate implicit knowledge of the disease, could be obtained by coupling a high value for the parameter associated with the salience of disease-related words with a low value of the salience of negative words (Figure 2).
Conclusions:
Our generative model can replicate participants' observed behavior, confirming the viability of the strategy of model inversion to estimate individual parameter values and illustrating the face validity of the model. These estimates can then be compared with MRI data, but also with neuropsychological test scores commonly used to assess anosognosia severity. By employing the control group as a baseline to account for non-disease-specific effects, this approach may facilitate the establishment of an implicit metacognitive measure in patients with anosognosia.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)
Emotion, Motivation and Social Neuroscience:
Emotional Perception
Modeling and Analysis Methods:
Bayesian Modeling 1
Perception, Attention and Motor Behavior:
Consciousness and Awareness 2
Keywords:
Aging
Meta-Cognition
Modeling
Statistical Methods
Other - Active-Infrence. Anosognosia; Alzheimer's
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
Mograbi, D. C., & Morris, R. G. (2013). Implicit awareness in anosognosia: Clinical observations, experimental evidence, and theoretical implications. Cognitive Neuroscience, 4(3–4), 181–197. https://doi.org/10.1080/17588928.2013.833899
Parr, T., Pezzulo, G., & Friston, K. J. (2022). Active Inference: The Free Energy Principle in Mind, Brain, and Behavior. The MIT Press. https://doi.org/10.7551/mitpress/12441.001.0001
Tondelli, M., Benuzzi, F., Ballotta, D., Molinari, M. A., Chiari, A., & Zamboni, G. (2022). Eliciting Implicit Awareness in Alzheimer’s Disease and Mild Cognitive Impairment: A Task-Based Functional MRI Study. Frontiers in Aging Neuroscience, 14, 816648. https://doi.org/10.3389/fnagi.2022.816648