A spiking neural network model of canonical babbling development

Abstract

The author wished to add the following correction:"After publication, it was discovered that the model did not receive reinforcement in the way that the paper described. The main conclusions still hold after correcting the error and making the following changes: performing synaptic weight normalization in place of spike-timing-dependent depression, removing the constant low level of dopamine, and scaling muscle activity by 2.5. A simulation and yoked control of the corrected model exhibited an increase in performance over time in the main simulation (r = 0.096, p < 0.001) as well as, to a lesser extent, in the control (r = 0.070, p = 0.008), and performance was significantly higher overall in the main simulation compared to control (T = 1.870, p = 0.031)."Abstract-Canonical babbling, the production of vocalizations that contain mature-sounding syllables, is one of the most striking and important milestones prior to the production of first words. This study simulates the emergence of canonical babbling using a spiking neural network containing motor neurons that activate muscles in a vocal tract simulator. The spiking neural network periodically produces synthesized vocalizations and a human listener judges the vocalizations on the basis of their syllabicity, deciding whether or not to reward the model. If a reward is given, spike timing dependent plasticity is increased and the model becomes more likely to recreate a pattern of neural firings similar to that which generated the reinforced vocalization. The model successfully increases its production of mature-sounding canonical syllables, whereas a yoked control simulation does not exhibit any such effect. This finding corresponds to results of experimental work with human infants in which consonantvowel syllable production is selectively reinforced by the infants' caregivers.

The Self-Organization of Human Interaction

“…Therefore, the oscillators must be entrained in antiphase, giving the participants both a common rhythm, constituted by speech rate and length of comfortable pauses, and complementarity-readiness to take the floor must be opposite at any given moment for speaker and hearer. This ability seems to appear at a very early developmental stage (Gratier & Devouche, 2011;Murray & Trevarthen, 1985;Nadel, Carchon, Kervella, Marcelli, & Réserbat-Plantey, 1999;Spurrett & Cowley, 2004;Warlaumont, 2012).…”

Emergent Jaw Predominance in Vocal Development Through Stochastic Optimization

“…Sensorimotor and social learning processes have also been used to explain rhythmic jaw movement. Warlaumont [41], [43], [42] proposes computational models of syllabic structure emergence based on social or intrinsic reinforcement. The model starts with random vocalisations produced using an articulatory synthesizer, i.e a computer model of the human vocal tract able to synthesize sound from articulatory movements.…”

“…Another line of work considers that rhythmic jaw movement can be the result of sensorimotor and social learning processes. Warlaumont [41], [43], [42] proposes computational models of syllabic structure emergence based on social or intrinsic reinforcement. Curiosity-driven self-exploration and imitation have also been considered as causal mechanisms to model the emergence of canonical babbling in speech acquisition [25].…”

Modeling the emergence of syllabic structure

“…At least two recent models, one by Moulin-Frier, Nguyen, and Oudeyer (2014) and one by myself (Warlaumont, 2012(Warlaumont, , 2013, have attempted to totally do away with the assumption of a pre-existing syllabic frame and instead explain how syllables containing both consonants and vowels could emerge through learning. Moulin-Frier et al's (2014) model controls vocalization by setting the locations and heights of the centers of five Gaussian functions, each defined in a multidimensional space the size of the number of articulators in the model's vocalization synthesizer.…”

“…I have utilized a recurrent spiking neural network to control movement of the masseter (which promotes jaw closure) and orbicularis oris (which promotes lip closure) during vocalization (Warlaumont, 2012(Warlaumont, , 2013. The model contains a network of 1000 spiking cortical neurons.…”