Research

Conversation depends on mutual understanding, emphasizing the need for an interactive alignment of meaning. When a conversation evokes personal and even idiosyncratic memories and associations we might choose to keep such thoughts private. A key question is whether interacting with others may not only affect what we say, but also what kind of thoughts come to mind, and how fluently they emerge. Some of the questions we investigate in this project are: How do social contexts affect our thought dynamics? What are the implications of failing to adapt, or indeed excessively adapting our thoughts to an interlocutor? How can impairments in these mechanisms explain different mental health problems, such as schizophrenia and social anxiety?

We all sometimes have unwanted thoughts we wish away. In severe cases, especially when such unwanted thoughts become repetitive, they can have a central role in the development or maintenance of different psychopathologies. Unfortunately, the scientific study of mechanisms driving unwanted repetitive thought has proven particularly challenging, partially due to their inherently covert nature. In this project, we focus on the idea that thoughts are self-amplifying: just like retrieving a memory strengthens it, thinking a thought makes it more likely to recur, and can eventually make it repetitive. We ask how this self-amplification effect comes about, and whether and how it can be regulated or mitigated to minimize repetitive thought 

Relevant papers:

Fradkin, I., & Eldar, E. (2022). If you don’t let it in, you don’t have to get it out: Thought preemption as a method to control unwanted thoughts. PLoS Computational Biology, 18(7).

Effective communication is at the heart of social integration and well-being. In psychopathology, incoherent discourse is often assumed to reflect disorganized thinking and is classically linked to psychotic disorders. However, people do not express everything that comes to mind, rendering inferences from discourse to the underlying structure of thought challenging. In fact, a range of psychopathologies are linked to self-reported disorganized thinking yet show little evidence of overt language incoherence. This project aims to investigate the mechanisms of incoherent discourse in psychosis and in other psychopathological dimensions, through a combination of computational modeling of free association and natural language processing.

Relevant papers:

Fradkin, I., Adams, R. A., Siegelman, N., Moran, R., Dolan, R. J. (2024). Latent Mechanisms of Language Disorganization Relate to Specific Dimensions of Psychopathology. Nature Mental Health.  

Natural language processing holds promise to transform psychiatric research and practice. A pertinent example is the success of natural language processing in the automatic detection of speech disorganization in formal thought disorder. However, we lack an understanding of precisely what common natural language processing metrics measure and how they relate to theoretical accounts of formal thought disorder. 

We tackle these questions by using deep generative language models to simulate incoherent narratives by perturbing computational parameters instantiating theory-based mechanisms of formal thought disorder.

Relevant papers:

Fradkin, I., Nour, M. M., & Dolan, R. J. (2023). Theory-driven analysis of natural language processing measures of thought disorder using generative language modeling. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 8(10)

The associative manner by which thoughts follow one another has intrigued scholars for decades. How do we efficiently `select` an association despite considerable competition from countless alternatives? By conceptualizing free association as the product of internal evidence accumulation, we can test and compare the computational principles that can explain the efficient generation of free association. One such principle is called Rich-get-richer dynamics, according to which an association that started to accumulate `evidence` is more likely to accumulate further evidence, effectively reducing competition over time. Another principle involved the Parallel spread of activation to all associations, in which case noise in the process enables mitigating the degree of competition among associations.

Relevant papers:

Fradkin, I., & Eldar, E. (2022). Accumulating evidence for myriad alternatives: Modeling the generation of free association. Psychological review. 

One of the most prominent qualities of people's mental experiences, such as thoughts, is that they are theirs alone. Perhaps the most extreme example of being alienated from one's thoughts is that of thought insertion, reported by some patients with schizophrenia. However, more subtle examples can be found in other conditions. People with OCD, for example, often describe their intrusive thoughts as `coming out of nowhere`, implying a problematic sense of agency over these thoughts. In this project, we investigated the mechanistic underpinning of the intrusiveness of thought, focusing on their predictability and sense of agency.

Relevant Papers:

Fradkin, I., Eitam, B., Strauss, A. Y., & Huppert, J. D. (2019). Thoughts as unexpected intruders: Context, obsessive-compulsive symptoms, and the sense of agency over thoughts. Clinical Psychological Science, 7(1), 162-180. 

Fradkin, I., & Huppert, J. D. (2018). When our train of thought goes off track: The different facets of out-of-context thoughts in obsessive compulsive disorder. Journal of obsessive-compulsive and related disorders, 18, 31-39. 

When locking your door, you rely not only on sensory information (seeing, hearing, and feeling a click), but also on a prediction that locking the door determines that it is locked and will remain that way unless someone unlocks it. In this project we explored the idea that obsessive-compusive symptoms, which include repetitive checking, washing, etc., results from an impairment in this prediction mechanism.

Relevant Papers:

Fradkin, I., Simpson, H. B., Dolan, R. J., & Huppert, J. D. (2023). How computational psychiatry can advance the understanding and treatment of obsessive‐compulsive disorder. World Psychiatry, 22(3).

Fradkin, I., Adams, R. A., Parr, T., Roiser, J. P., & Huppert, J. D. (2020). Searching for an anchor in an unpredictable world: A computational model of obsessive-compulsive disorder. Psychological Review, 127(5).

Fradkin, I., Ludwig, C., Eldar, E., & Huppert, J. D. (2020). Doubting what you already know: Uncertainty regarding state transitions is associated with obsessive compulsive symptoms. PLoS Computational Biology, 16(2).