| A heated argument in visual word recognition studies is whether it involves cascaded processing without any thresholds in a highly interactive system reflected in the Interactive Activation Model or the discrete, sequential processing in the separate modules. According to the classical Sternberg’s additive-factors logic, if two variables which are manipulated in one experiment exert no interactive effects, namely additive effects, it suggests that the two variables influence separate processing stages, however, if they are found to be interactive effects, it means they influence at least one stage in common. According to the Interactive Activation Model, the different modules were nearly synchronous, namely a common stage, so the joint effect of stimulus quality and word frequency was interactive. However, for decades there was a conundrum confusing the supporters for the Interactive Activation Model:the observation that stimulus quality and word frequency always produced additive effects. This pattern may be more applied for the discrete processing embodied in the multistage activation model. According to the assumption, stimulus quality that had an effect on the letter level, and word frequency which affected the pathway that transmitted the activated representations from the former stage to their corresponding representations in the following stage rather than the stage itself should be additive for respectively affecting the different parts of the lexical processing. Whereas, this phenomenon may result from the phonology processing as an "obstacle" made the connections between the orthographic processing and semantic processing weak in alphabetic language with reducing the possibility of the interaction of the stimulus quality and frequency. Then by contrast, Chinese characters can directly activate their semantic representations and showed a greater reliance on orthography. As a result, there should be an interactive effect between the stimulus quality and frequency in Chinese characters on N400 which was an index of semantic processing. The current study aimed to adopt a purer paradigm, the delayed character-matching task without any priming, to perceive the joint effect of the stimulus quality and word frequency on N400 referred to the Sternberg’s additive-factors logic. It seemed to be an effective methods to evaluate the suitability of the two frameworks or the processing mode.The experiment 1 was a 2 (stimulus quality:highly-blurred vs. slightly-blurred) x2 (frequency: high vs. low) within-subjects design. The task presented the target character which was marked in the current ERP first, then the probe either matched with the target or mismatched. The participants were asked to press buttons to judge whether they were in identity regardless of different shapes. The results of behavioral data showed a significant main effect of stimulus quality with the slower RT in the response of highly-blurred stimuli compared with slightly-blurred ones, and a significant main effect of frequency with slower RT in the response of low frequency characters than high frequency ones. Meanwhile, there was a significant interactive effect of the two variables on RT. The simple effect analysis of the interaction demonstrated the RT was significantly slower in the highly-blurred condition than the slightly-blurred one during the low frequency, but not during the high frequency. It could be also seen a significant main effect of stimulus quality with the higher ACC in the slightly-blurred condition and a significant main effect of frequency with the higher ACC of high frequency, but no interaction. The ERP data showed that in an N400 time window of 400-500 ms, there was a significant three-way interactive effect of stimulus quality, frequency and electrodes. The simple effect analyses of the three-way interactive effect exhibited a significant two-way interactive effect of stimulus quality and frequency at right central-parietal site. The simple effect analyses of the two-way interactive effect displayed the average amplitude of N400 following the highly-blurred stimuli were bigger than the slightly-blurred ones when the frequency of targets were low, but not in high frequency. There was an interaction of the stimulus quality and frequency on both RT and N400 demonstrated that the mode of cascaded processing applied to Chinese characters recognition and just provided evidence for the Interactive Activation Model, however, made a suspicion of the multistage activation model. In addition, in the time window of 220-280 ms, the highly-blurred stimuli elicited a larger N250 than the slightly-blurred stimuli regardless of the level of frequency. The N250 stimulus quality effect may be the result of mismatching between the mental template to a Chinese character in our brains and the blurred form in the present task.The experiment 2 aimed to acquire a stricter interaction of stimulus quality and frequency based on experiment 1. The stimulus quality was operated as the level of consciously unidentifiable blurredness to generate unconscious processing. It was a 2 (stimulus quality:unidentifiable-blurred vs. identifiable-blurred) ×2 (frequency:high vs. low) within-subjects design. The results of behavioral data showed a significant main effect of stimulus quality with the slower RT in the response of highhly-blurred stimuli compared with slightly-blurred ones. Meanwhile, it could be also seen significant main effects of stimulus quality and frequency with the higher ACC in the identifiable-blurred condition and high frequency, but no interaction. The ERP data showed that in an N400 time window of 550-750 ms which was a little late because the targets were unidentifiable-blurred, there was a significant two-way interactive effect of stimulus quality and frequency. The simple effect analyses of the interaction displayed the average amplitude of N400 following the unidentifiable-blurred stimuli were bigger than the identifiable-blurred ones when the frequency of targets were low, but the difference was small in high frequency. The results further proved the cascaded processing in Chinese. Meanwhile, there was an N250 stimulus quality effect in the time window of 200-300 ms referred to that the average amplitude of N250 was larger in the condition of the unidentifiable-blurred stimuli than the slightly-blurred ones. This further proved the conclusion in experiment 1.In conclusion, firstly, the interaction of the stimulus quality and frequency on N400 in the two experiments was quite different from the previous studies where the joint effect of the two factors was always additive in alphabetic language. The present study demonstrated that the mode of cascaded processing applied to Chinese characters recognition. Because there was a strong link between orthographic and semantic processing in Chinese, Chinese characters could directly activate their semantic representations through the orthographic information and made the mode of cascaded processing possible. The results provided evidence for the the Interactive Activation Model, however, made a suspicion of the multistage activation model.Secondly, there was an N400 stimulus quality effect in low frequency suggested that the stimulus quality could modulate the N400 amplitude by influencing the semantic processing directly with the less pre-activation like the highly-blurred, the unidentifiable-blurred stimuli, the more additional activation, so as to the larger N400. This finding supported the N400 notion of semantic retrieval theory.Thirdly, the N250 stimulus quality effect that was referred to the highly-blurred, the unidentifiable-blurred stimuli elicited larger N250 than the slightly-blurred, identifiable-blurred ones regardless of the frequency indicated that N250 was sensitive to the the number of mismatching letters between prime and target, in order words, N250 maybe was an index of orthographic processing in Chinese characters. |
PDF Full Text Request