Visual working memory (VWM) is characterized by strict limitations in capacity and precision, whose nature remains the subject of an ongoing theoretical debate between discrete-representation models and continuous-resource models. The present study investigates the mechanisms underlying VWM using a delayed estimation task (Color Wheel Task) administered to a sample of healthy adults, while systematically manipulating memory set size. Behavioral data were analyzed using a three-parameter mixture model estimated within a hierarchical Bayesian framework and subsequently fitted to three competing computational models: Slot-plus-Averaging (SAF), Equal Precision Fixed (EPF), and Variable Precision (VPA). Model comparison, conducted using information criteria (AIC) and predictive error measures (RMSE), revealed that the variable precision model provided the best fit to the data. Overall, the findings support a view of visual working memory as a system based on a continuous and stochastically variable resource rather than a fixed number of discrete representations.
La memoria di lavoro visuo-spaziale (Visual Working Memory, VWM) è caratterizzata da limiti stringenti di capacità e precisione, la cui natura è oggetto di un ampio dibattito teorico tra modelli a rappresentazione discreta e modelli a risorsa continua. Il presente studio indaga i meccanismi sottostanti la VWM mediante un compito di delayed estimation (Color Wheel Task) somministrato a un campione di adulti sani, manipolando il numero di elementi da memorizzare. I dati comportamentali sono stati analizzati attraverso un three-parameter mixture model stimato in un framework gerarchico bayesiano e successivamente adattati a tre modelli computazionali concorrenti: Slot-plus-Averaging (SAF), Equal Precision Fixed (EPF) e Variable Precision (VPA). Il confronto tra modelli, condotto mediante criteri informativi (AIC) e misure di errore predittivo (RMSE), indica che il modello a precisione variabile fornisce il miglior adattamento ai dati. I risultati supportano una concezione della memoria di lavoro visuo-spaziale come sistema basato su una risorsa continua e stocasticamente variabile, più che su un numero fisso di rappresentazioni discrete.
Modelli computazionali della memoria di lavoro visuo-spaziale: un’indagine sperimentale mediante il Color Wheel Task
BERTOLINI, GIANMARCO
2025/2026
Abstract
Visual working memory (VWM) is characterized by strict limitations in capacity and precision, whose nature remains the subject of an ongoing theoretical debate between discrete-representation models and continuous-resource models. The present study investigates the mechanisms underlying VWM using a delayed estimation task (Color Wheel Task) administered to a sample of healthy adults, while systematically manipulating memory set size. Behavioral data were analyzed using a three-parameter mixture model estimated within a hierarchical Bayesian framework and subsequently fitted to three competing computational models: Slot-plus-Averaging (SAF), Equal Precision Fixed (EPF), and Variable Precision (VPA). Model comparison, conducted using information criteria (AIC) and predictive error measures (RMSE), revealed that the variable precision model provided the best fit to the data. Overall, the findings support a view of visual working memory as a system based on a continuous and stochastically variable resource rather than a fixed number of discrete representations.| File | Dimensione | Formato | |
|---|---|---|---|
|
Bertolini_Gianmarco.pdf
accesso aperto
Dimensione
2.3 MB
Formato
Adobe PDF
|
2.3 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/107817