Pregled bibliografske jedinice broj: 1150845
Neurodynamical Models of Embodied Cognition
Neurodynamical Models of Embodied Cognition // Knjiga sažetaka / Savremeni trendovi u psihologiji, Filozofski fakultet, Novi Sad, 19-21. oktobar 2017. = Book of abstracts / Current Trends in Psychology, Faculty of Phiosophy, Novi Sad, October 19st-21st, 2017. - Novi Sad : Filozofski fakultet, 2017
Novi Sad: Filozofski fakultet Univerziteta u Novom Sadu, 2017. str. 24-25 (pozvano predavanje, međunarodna recenzija, sažetak, znanstveni)
CROSBI ID: 1150845 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Neurodynamical Models of Embodied Cognition
Autori
Domijan, Dražen ; Šetić, Mia
Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni
Izvornik
Knjiga sažetaka / Savremeni trendovi u psihologiji, Filozofski fakultet, Novi Sad, 19-21. oktobar 2017. = Book of abstracts / Current Trends in Psychology, Faculty of Phiosophy, Novi Sad, October 19st-21st, 2017. - Novi Sad : Filozofski fakultet, 2017
/ - Novi Sad : Filozofski fakultet Univerziteta u Novom Sadu, 2017, 24-25
ISBN
978-86-6065-434-4
Skup
Savremeni trendovi u psihologiji 2017
Mjesto i datum
Novi Sad, Srbija, 19.10.2017. - 21.10.2017
Vrsta sudjelovanja
Pozvano predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
grounded cognition ; perception ; motor control ; neural model
Sažetak
Research on grounded cognition suggests that the processing of a word or concept reactivates the perceptual and/or motor representations associated with the referent object. Such reactivation then interacts with the ongoing cognitive computation. It is interesting to note that the effects of motor or online grounding are generally stronger and more robust when compared to perceptual or offline grounding. The aim of this study is to show how the theory of complementary brain processes explains differences between online and offline grounding as manifestations of different cortical circuits designed to solve complementary tasks (Grossberg, 2000 ; 2009). We suggest that perceptual grounding arise from the neural network designed to achieve stable category learning as elaborated in the adaptive resonance theory (ART). It affords fast learning without catastrophic forgetting by comparing sensory (bottom-up) data with learned (top-down) expectations. If the input pattern matches with one of the previously learned codes (categories) it is recognized as a familiar pattern. On the other hand, if there is a mismatch between the input pattern and the learned code, a reset signal is triggered that removes the currently active category representation and initiates a search for another category that may provide better match with the input. Previously, we suggested that the same top-down pathway that is needed to stabilize learning is also used during conceptual processing to reactivate perceptual patterns associated with concept (Domijan & Šetić, 2016). We showed that a mismatch between top-down expectation and bottom-up sensory data activates a reset signal that slows execution of the current cognitive task thus producing interference effect in incongruent condition (i.e., in condition where sensory data such as perception of blue color do not agree with the expectation produced by the concept such as word RED). Also, we showed that congruent condition does not produce any effect because top-down pathway is modulatory and cannot directly activate sensory nodes without direct sensory stimulation. Here, we demonstrated that online or motor grounding arises from the neural circuit designed to control arm movement. Such circuit does not need reset signal to correct its errors and top-down pathway to its nodes does not need to be modulatory. Therefore, words and concepts that activate motor plans will produce interference in incongruent as well as facilitation in congruent experimental condition. Consequently, online grounding will be more robust and stronger when compared to offline grounding which exhibits only interference. We conclude with the suggestion that further progress in studying grounded cognition will benefit from considering established neurocomputational mechanisms that can support such complex cognitive phenomena.
Izvorni jezik
Engleski
Znanstvena područja
Psihologija, Kognitivna znanost (prirodne, tehničke, biomedicina i zdravstvo, društvene i humanističke znanosti)
POVEZANOST RADA
Ustanove:
Filozofski fakultet, Rijeka,
Hrvatsko katoličko sveučilište, Zagreb