Primate brain is highly capable to integrate the information of multiple sensory areas together to perceive environment uniquely. In a particular sensory system, this integration of features known as “feature binding”. In the visual system, various visual features of an object thought to be dominantly processed in the separate areas of the brain. For instance, color in V4, motion direction in MT, shape in V1 and V2 etc. The main question of visual feature binding is how the brain integrates these features of an object together and separates them from other objects.
Two main theories related to feature binding are “binding by synchrony” (BBS) and “feature integration theory” (FIT). The BBS theory proposed that feature binding is performed by synchronous activity of the involving neurons (Von der Malsburg, 1994, Elliott& Müller, 1998, Gray and Singer, 1989). However, it is limited to this fact that how two components of the same dimension could be bounded (grouping), rather than two features from different dimensions (Shadlen & Movshon, 1999 Thiele & Stoner, 2003, Palanca & DeAngelis, 2005, Ray, 2010, Martin, Anne, 2015). In FIT theory the perception of an object is completed in two stages, feature separation, and feature combination (Treisman and Gelade, 1980). This theory suggests the attention as the main component in feature integration mechanism (Treisman and Gelade, 1980, Treisman & Schmidt 1982, Schoenfeld 2003, Cohen. 2013, Katzner 2006, Golomb, L’Heureux, & Kanwisher, 2014, Robertson, 2003). A number of studies reported some forms of feature conjunction in absent of attention (Houck 1986, Gajewski, 2006, Mordkoff, 2008) or similar function of attention in working memory (Allen, Hitch, & Baddeley, 2009; Brown & Brockmole, 2010; Fougnie & Marois, 2009).
The feature binding has been reported in several areas of the brain; from early visual areas (Seymour et al, 2009) to parietal cortex (Friedman, 1995, Shafritz, 2002). Overall, the main mechanism of feature integration remains a matter of debate.Furthermore, since most studies of visual feature binding were carried out based on functional imaging and psychophysics, a lack of neural evidence of this mechanism is observed.
A clue to resolve the binding problem is the probability of controlling this mechanism by a higher-level area of brain. It is widely believed that cortex encodes visual stimuli in a hierarchical manner (Riesenhuber and Poggio, 1999). Prefrontal cortex is an ideal region for this purpose. It houses the information from both ventral and dorsal visual streams (SFuster, 1988). In addition, single neurons of this area encodes the diection, color, position and behaviorally relevence visual features (Rainer et al., 1998; Ninokura et al., 2004, Miller and Cohen, 2001; Rigotti et al., 2013). Furthermore, it had been revealed that there is a task-specific top-down information transmission from prefrontal cortex to early visual areas (Yosuke Morishima 2009).
To examine the role of PFC in feature binding, using an electrophysiological study we investigated about the existence of feature conjunction in the prefrontal cortex. Results indicate that the integration of color and motion direction features, which are dominantly emergenced in the areas v4 and MT respectively, encodes in the prefrontal cortex. Tis coding arrises in the low-frequency theta and alpha of LFP signal. We further, found that the feature binding emergences in the single sites of PFC neurons and propagates into surronding neurons. Moreover, we showed that this binding information is behaviorally relevant.