Neural correlates of the food/non-food visual distinction

Distinct brain representations of processed and unprocessed foods

“…Greater activation in visual areas has been observed, in response, for instance, to high-calorie foods (Killgore et al, 2003;Simmons, Martin, & Barsalou, 2005;Toepel et al, 2010), as well as to other highly relevant stimuli such as monetary rewards (Small et al, 2005) or emotional faces (Eger, Jedynak, Iwaki, & Skrandies, 2003;Moratti, Keil, & Stolarova, 2004). The difference in occipital activation observed in the present research, in line with previous literature with food stimuli (as in Stingl et al, 2010;Meule et al, 2013;Tsourides et al, 2016), is unlikely to be due to differences in visual complexity of the stimuli.…”

“…This early unprocessed-processed discrimination effect falls within the window of visual object recognition when other salient properties are also detected (Antal et al, 2000;Proverbio et al, 2007;Thorpe et al, 1996). Previous studies showed that the information regarding objects' edibility is coded by the brain at about 85 ms post-stimulus onset (Tsourides et al, 2016), objects' naturalness at about 120 ms (Cichy et al, 2014), and energy level at about 160 ms (Meule et al, 2013;Toepel et al, 2009). We argue that the within-category food unprocessed-processed discrimination is an instance of the broader between-category natural-artifact discrimination.…”

“…That higher-level effects may occur early in the timeline of object visual processing found confirmation in recent MEG studies. For instance, Tsourides et al (2016) provided evidence of a discrimination effect based on edibility in the ventral visual stream within 50 ms from the arrival of information in V1, as early as 85 ms post-stimulus onset stimuli. Likewise, Cichy et al (2014) reported that in the ventral visual cortex objects were discriminated for naturalness (natural vs. artificial), with a peak at 122 ms post-stimulus onset.…”

“…To date, a handful of electrophysiological studies have provided insight about the types of information that are extracted during visual object recognition. For instance, already around 85 ms the brain is able to discriminate between edible and non-edible items (Tsourides et al, 2016), while between 100-200 ms post-image onset it detects differences in processing visually presented high-and low-energy food (Lietti, Murray, Hudry, le Coutre, & Toepel, 2012;Meule, Kübler, & Blechert, 2013;Toepel, Knebel, Hudry, le Coutre, & Murray, 2009); around the same time (100-200 ms), the brain distinguishes between natural items and artifacts, animate and inanimate items (Antal, Keri, Kovacs, Janka, & Benedek, 2000;Cichy, Pantazis, & Oliva, 2014;Proverbio, Del Zotto, & Zani, 2007;Thorpe, Fize, & Marlot, 1996).…”

Deep neural networks on graph signals for brain imaging analysis

“…Therefore, MEG are high-dimensional spatiotemporal data often degraded by complex, non-Gaussian noise. For reliable analysis of MEG data, it is important to learn discriminative, lowdimensional intrinsic representation of the recorded data [1,2].…”