Naslov
A large-scale investigation of structural brain correlates of metacognition

Autori
Rutiku, Renate ; Juras, Luka ; Mazor, Matan ; Derda, Monika ; Filevich, Elisa ; Fleming, Steve ; Wierzchoń, Michał ; Sandberg, Kristian

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Skup
26th annual meeting of the Association for the Scientific Study of Consciousness

Mjesto i datum
New York (NY), Sjedinjene Američke Države, 22.06.2023. - 25.06.2023

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
metacognition ; quantitative multi-parameter mapping ; diffusion-weighted imaging

Sažetak
Despite the central role of metacognition in many theories of consciousness, we still know relatively little about its internal structure and its neural underpinnings. Previous attempts to extract domain-general behavioral indices of metacognition and map them to common neural sources have yielded mixed results. Thus it is not clear yet whether metacognition is a general trait-like ability or whether it applies differentially to specific tasks. We hypothesize that the previous mixed results are in large part due to small sample sizes and a lack of attention to the stability of the behavioral indices of metacognition (or indeed the metacognitive processes themselves). We aimed to address these shortcomings in a large- scale collaborative effort as part of the SkuldNet consortium/EU COST Action CA18106. Metacognitive ability was assessed in two independent samples (N1=301 ; N2=246), along with a comprehensive mapping of participants’ brain structure. Participants in the first cohort completed visual and auditory versions of a simple 2AFC discrimination task and a 2AFC detection task (4 tasks in total) and rated their confidence in response accuracy on every trial. A subsample of 54 participants also repeated the experiment on a separate day in order to determine the stability of metacognition over time. First, various common measures of metacognition were compared in terms of their stability within tasks and testing sessions, and between tasks and testing sessions. We will present the results of these comparisons and discuss potential reasons why some measures are more stable than others. For example, response bias is a strikingly stable behavioral marker of interindividual differences across tasks and time. Other measures of metacognition such as meta-d’ are significantly less stable. Second, we tested the relation between various behavioral indices and multimodal metrics of individual brain structure in order to explore the neural architecture underlying metacognition. Specifically, we used quantitative multi-parameter mapping (MPM) and diffusion-weighted imaging (DWI) to conduct a brain-wide association analysis and ROI analysis based on previously published findings. The results highlight how important it is to attempt replication of brain-behavior associations in larger cohorts because many of the findings cannot be verified. Yet accounting for the stability of the behavioral measures in the analysis may increase the robustness of the results. We will summarize which of the previous findings were replicated in our first cohort. We will also present results from an out-of-sample replication of our findings in our second cohort. Together, these findings highlight the complex multidimensional nature of metacognition and demonstrate how interindividual differences in behavior can be exploited to reveal its underlying neural architecture.

Izvorni jezik
Engleski

Znanstvena područja
Psihologija

POVEZANOST RADA