Bridging The "Gap" Between N170 And Recognition Potential

Object recognition, including reading, can be extremely rapid. Important computations occur within 300ms of seeing an object. Understanding those computations using ERP is critical for understanding the early information-sensitivity of object recognition. Current research mainly focuses on two well-known ERP waves:N170 and Recognition Potential(RP). N170 and RP are psychophysiological responses occurring at ～170ms and ～250ms; both of them are sensitive to words. The RP is usually obtained by a rapid stream stimulation procedure (RSS) (Rudell,1992), while N170 can be observed with a single presentation of visual stimuli. Studies of the RP often mention that RSS can enhance the signal-to-noise ratio (SNR). However, surprisingly after decades of research, none of studies has ever proved that RSS provides better signal than the N170 paradigm. Our primary aim is to finally determine if that is true. However, we must first determine if the N170 and RP signals are, in fact, comparable. The N170 and RP share similarities though their latencies are different. There some suspicions on whether they really represent two distinct stages of word recognition. Therefore, we should confirm that the N170 and RP reflect the same neural mechanisms.Although the traditional paradigm for the N170 and RP are distinct, these two potentials share several properties, e.g. location and stimulus sensitivity (Bentin et al.,1999; Marti’n-Loeches et al.,2005). Therefore, we suspect that they are same ERP and the longer latency of RP resulted from the forward mask, which can delay the psychophysiological response to the target. To examine this hypothesis, we adopted the revised paradigm of RSS, and manipulated the ISI (or gap) between the forward mask and the target. This experiment provides us with a range of conditions along a continuum between the N170 paradigm (no forward mask), and the RP paradigm (a mask immediately before the target). If the N170 and RP responses are the same component, we predict a gradual shift from an early N170 to a late RP. Chinese characters were used as targets and scrambled characters were used as masks. Current results confirmed our main hypothesis:for each subject, a gradual shift from the N170 to the RP as gap size decreased. Specifically, the speed of RP latency increased with gap-size, and approached N170-speed as an exponential decay function. Furthermore, individual differences in exponential decay were very well predicted by delays in the forward mask N1.Critically, the character-evoked N1s have similar topographies across gap conditions. And finally, variations in character-evoked Nl latency and amplitude are well-predicted by the vertex positive potential. Our results demonstrate that N170 and RP share the same underlying neural mechanism.Experiment 2 compared the inversion effect of face and Chinese characters under the RSS paradigm and the N170 paradigm. Our findings show that the effect is larger in the RSS paradigm than in the N170 paradigm. And the effect in the RSS paradigm is consistent with the previous N170 studies. Furthermore, the topography results show that the topographies of RP are highly correlated with the topographies of N170, which further confirms that RP and N170 share same underlying neural mechanism.In sum, our research has provided sufficient evidences to prove RP and N170 share same underlying neural mechanism and provided the experimental evidences that better signal noise ratio can be obtained with RSS paradigm. Our results also suggest that N170 and RP may relate to same stage of word recognition. Moreover, RSS can be used as a better tool in N170 studies.