Generic neural computational construct for instantaneous estimation: a pilot study

991 

Abstract Submission 

Lydia Yingzhe Li¹, Guang Ouyang²

¹The University of hong Kong, Hong Kong, Hong Kong, ²The University of Hong Kong, Hong Kong, Hong Kong

Lydia Yingzhe Li  
The University of hong Kong
Hong Kong, Hong Kong

Guang Ouyang  
The University of Hong Kong
Hong Kong, Hong Kong

Instantaneous estimation of scale, spanning numerical, spatial, and temporal domains, stands as a fundamental skill crucial to our daily functioning. For instance, the split-second estimations involved in throwing a grenade--calculating target distance, grenade weight, and explosion timing-signify its vital role in decisive moments. This efficiency hints at a shared computational construct underlying these estimations, which is also supported by empirical evidence such as SNARC effect (Dehaene et al., 1993).
However, a comprehensive investigation into a unified framework encompassing instantaneous estimations across various domains is lacking. Our study utilizes EEG's high temporal resolution to investigate these fast time scale cognitive processes. Both neural and behavioral indicators were collected to model the construct structure based on structure equation modelling. We aim to address two questions: 1. Are these estimations governed by unified brain computational mechanism? 2. If so, is it differentiable from other basic cognitive abilities such as working memory and mental speed?

We employed two distinct approaches based on five tasks: one examines the convergent validity by identifying shared construct across the three domains for instantaneous estimations, while the other examines the discriminant validity by comparing these constructs against working memory and mental speed (Figure 1).

Numerical estimation
Subtask 1 involves judging "more" or "less" on two dot sets that slightly vary in total number. Subtask 2 (control task) presents two frames: the first replicates subtask 1, while the second might add or remove five dots from the first frame. Participants judged if the two frames are identical. Visual consistency between subtask 1 and 2 enables us to isolate ERP effect related to numerical estimation.

Spatial estimation
Participants assessed whether a line passes through the center of a circle. Trials vary in difficulty based on the line-center distance. ERP differences between easy and hard conditions can be obtained to represent spatial estimation.

Temporal estimation
Subtask 1 involved discriminating between sounds of slightly different durations. Subtask 2 (control task) presented the same sounds without prompting duration estimation. Auditory consistency between two subtasks allows us to isolate ERP effect related to temporal estimation.

Working memory
Delayed match-to-sample paradigm was adopted. Participants identified if the three delayed comparison stimuli are identical with the target. We assessed the working memory load by comparing ERPs corresponding to the sample and comparison stimuli.

Mental speed
This task measures how quickly participants respond to stimuli by pressing a button. We examined the N1 latency and used it as a neural indicator of perceptual mental speed.

Based on 15 participants in the pilot study, we have successfully identified significant neural effects from all of tasks (except for the mental speed where only latency is needed. See the detailed effect patterns in Figure 2A-E). These neural effects were used to examine the inter-task relationships, shown in the correlation matrix in Figure 2F (together with the behavioral counterpart). The correlations between numerical, spatial, and temporal estimation largely aligned with our expectations. We have also conducted power analysis based on the correlation (Wang & Rhemtulla, 2020), which estimated a requirement of 200 participants. The structural relationships between the construct of instantaneous estimation and other basic abilities will be examined by SEM as shown in Figure 2G.

Our study on instantaneous estimation across domains unveils intricate task relationships. Initial findings show intriguing connections despite some divergence, highlighting the necessity for further investigation using a larger sample via structural equation modeling.

Space, Time and Number Coding ¹

EEG ²

Cognition

Electroencephaolography (EEG)

Experimental Design

Other - instantaneous estimation ability; event related potential (ERP)