Odor naming is considered as a particularly challenging task. In free odor naming scenarios people often fail to respond with any linguistic label to certain odors they smell, resulting in what is called an omission (i.e., a lack of response). The cognitive demands and the nature of interactions between olfaction and language related brain’s neural resources are not well understood [1]. Behavioral task: the data that we used to compare our model output were collected from a Swedish National Study on Aging and Care in Kungsholmen (SNAC-K), where ~2500 subjects underwent an odor naming test of 16 odors [2]. After presenting each individual odor, the SNAC-K participants were asked to freely identify and name the presented stimulus (odor naming). Μemory model and Results: with the support of a computational model we offer a hypothesis about neural network mechanisms underlying the frequently observed phenomenon of omissions in free odor naming tasks. We used distributed network representations for the odor (percept) and word label representations, and attempted to account for their statistical inter-relationships (correlations) extracted from the available data on odor perceptual similarity from a Swedish odor language corpus, respectively [3,4,5]. Our meta-analysis on the available behavioral data [2] suggests that odors with numerous language associations (one-to-many mapping) lead to elevated blank responses (Fig. 1B, SNAC-K data). Using our memory model (Fig. 1A) consisting of two networks that are reciprocally connected with Bayesian-Hebbian plasticity [6], we first simulated the memory task, and reproduced the trends observed in the SNAC-K behavioral data (Fig. 1B, Model). Further, we analyzed changes in the network's synaptic plasticity to provide a mechanistic explanation for the elevated omissions in the one-to-many mapping scenario. Due to the nature of Bayesian-Hebbian associative learning of the synaptic weights connecting the two networks, there was a weak coupling for odors paired with multiple different labels during the odor-label associative pre-encoding (Fig. 1C), thereby resulting in subthreshold olfactory language network responses (omissions) corresponding to those perceptual odor stimuli. Additionally, our model predicts that poor performance in odor naming tasks may arise from a high overlap between the neural representations of different odor percepts (Fig. 1D), which often leads to failure in recognizing the odor.