The Evolution of Meaning: Spatio-temporal Dynamics of Visual Object Recognition

Abstract

Abstract■ Research on the spatio-temporal dynamics of visual object recognition suggests a recurrent, interactive model whereby an initial feedforward sweep through the ventral stream to prefrontal cortex is followed by recurrent interactions. However, critical questions remain regarding the factors that mediate the degree of recurrent interactions necessary for meaningful object recognition. The novel prediction we test here is that recurrent interactivity is driven by increasing semantic integration demands as defined by the complexity of semantic information required by the task and driven by the stimuli. To test this prediction, we recorded magnetoencephalography data while participants named living and nonliving objects during two naming tasks. We found that the spatio-temporal dynamics of neural activity were modulated by the level of semantic integration required. Specifically, source reconstructed time courses and phase synchronization measures showed increased recurrent interactions as a function of semantic integration demands. These findings demonstrate that the cortical dynamics of object processing are modulated by the complexity of semantic information required from the visual input. ■

Interactions between visual and semantic processing during object recognition revealed by modulatory effects of age of acquisition

“…An alternative possibility is that differences in the quality and richness of early-compared with later-acquired semantic representations (Belke et al, 2005;Brysbaert et al, 2000;Steyvers and Tenenbaum, 2005) are reflected in activation of greater numbers of anterior temporal neurons and synapses, resulting in a stronger neuromagnetic response. In either case we note that the time period in which we observed modulation of the semantic response in left anterior temporal cortex by AoA overlaps with the period in which Clarke et al (2011Clarke et al ( , 2013 reported stronger activation of medial left anterior temporal cortex by basic than domain level naming and a modulation of responses at the same site by distinctiveness. It also overlaps with period that Bemis and Pylkkänen (2011) associated with the process of integrating the meanings of pairs of words.…”

“…Comparison of responses to pseudohomophones and orthographic controls in the left posterior IFG/precentral gyrus found a stronger response to targets following pseudohomophone than orthographic control primes around 100 ms (30-40 Hz) that was in the induced (non-phase-locked) rather than the evoked (phase-locked) component, with additional modulations between 200 and 500 ms in both the evoked and the induced components. Clarke et al (2011) used MEG data from a study of object naming to argue for on-line interactions between posterior visual and anterior semantic processing. Participants named pictures of objects at either a domain level (saying "living" or "manmade" in response to the pictures) or at a basic level (e.g., saying "cow" or "hammer").…”

“…When Bookheimer et al (2005) compared overt with covert object naming, they found very similar activation patterns except in regions associated with motor activity. Some previous MEG studies of object naming have used overt naming (Clarke et al, 2011(Clarke et al, , 2013Sörös et al, 2003;Vihla et al, 2006) while others have used passive viewing (Maratos et al, 2007), covert naming (Liljeström et al, 2009) or delayed overt naming in an effort to minimize mouth movement artifacts. Covert naming has been used in studies of visual word recognition for similar reasons (e.g., Wheat et al, 2010;Woodhead et al, 2011).…”

“…That said, Laaksonen et al (2012) reported induced ("rhythmic") activity at both 7-12 Hz and 17-22 Hz in left superior temporal regions during overt and covert object naming that extended down to the temporal pole. Clarke et al (2011) reported an evoked response from left medial anterior temporal cortex that was stronger during basic than domain level object naming from ~ 125 to 300 ms. Clarke et al (2013) left anterior temporal region from 84 to 120 ms and again from 170 to 210 ms. We also note that Bemis and Pylkkänen (2011) reported a period from ~ 185 to 255 ms during which MEG activity in left anterior temporal cortex was associated with combining the meanings of two words (e.g., red and boat) to decide whether or not they matched a simultaLog name frequency (Celex) (Baayen et al, 1993) Log name frequency (WFG) (Zeno et al, 1995) neously presented picture. Taken together, these studies indicate that it is possible to detect MEG responses originating in anterior temporal cortex and that those responses may be sensitive to semantic factors and semantic manipulations within the first 300 ms following stimulus presentation.…”

“…If the modulation of occipital activity by AoA observed by Ellis et al (2006) is a result of top-down re-activation of visual cortex by semantic processing (Clarke et al, 2011(Clarke et al, , 2013, then the modulatory effects of AoA on occipital activity would be expected to follow the modulatory effects on anterior temporal activity.…”