The neuropsychological basis of the Chinese mental lexicon: An ERP study using LORETA
Sheng Xi He,Wen Gang Yin,Yue Jia Luo,Brendan Stuart Weekes
DOI: https://doi.org/10.1016/j.bandl.2006.06.109
IF: 2.781
2006-01-01
Brain and Language
Abstract:Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Graduate school of CAS, Beijing 100101, ChinaAccepted 6 July 2006IntroductionThe aim of this study was to investigate the event-related potentialsinvolved in the processing of orthography, phonology and meaning duringChinese character recognition. Event-related potentials (ERPs) are thesummed activity of synchronized post-synaptic EEG potentials recordedon the scalp. ERPs can illuminate the temporal dynamics of the cognitiveactivity underlying lexical access and semantic integration in word recog-nition because of excellent temporal resolution (Curran, Tucker, & Kutas,1993). The neural sources of each ERP component can also be localizedwith low-resolution electromagnetic tomography (LORETA). Topo-graphic mapping is a means to display electrical activity in the brain onthe basis of multi-channel EEG recordings using temporal resolution thatis recorded in milliseconds. LORETA is an algorithm that solves theinverse problem of EEG and is performed on averaged ERPs to isolatethe spatial distribution of the components by finding the smoothest ofall possible solutions consistent with the scalp distribution (Pascual-Mar-qui, Michel, & Lehmann, 1994). The LORETA procedure calculates cur-rent densities in 2394 pre-selected brain voxels based on the scalpdistribution of potentials. The spatial reference used for this procedureis the Talairach brain atlas (Talairach & Tournoux, 1988). The spatial res-olution is 7 mm and the maximal error is within 10 mm (Pascual-Marqui,1999). Topographic mapping is shown as a distribution of colour-codedvalues on the head surface. The head surface is used to associate surfacenormal orientations to electrodes and also to compute radial and tangen-tial current density components. The distribution of topographic valuescan be interpreted as the distribution of the local radial current densityfrom just underneath the skull layer. Thus, topographic mapping givesan approximate image of cortical processes. If positive and negative topo-graphic values of high magnitude are found in any one scalp area, it can beconcluded that superficial brain activity occurred under this area (Pascual-Marqui, 1999). Very few studies have used EEG methods to examine Chi-nese character recognition (Liu, Perfetti, & Hart, 2003; Valdes-Sosa et al.,1993). All models of lexical access in Chinese (e.g. Perfetti, Liu, & Tan,2005) assume three independent cognitive stages—orthographic, phono-logical and semantic. In the present study, we used LORETA to examinetopographic values associated with these cognitive processing stages thatwe assume are necessary in character identification. We predicted thattasks designed toisolate theseprocesses would activate differentERP com-ponents and brain regions for orthographic, phonological and semanticprocessing.Participants and methodsTwenty right-handed native Chinese speaking students from the Agri-cultural University of China (10 males and 10 females) aged between 20and 23 years (Mean=22, SD=1.12) with no history of neurological or psy-chiatric illness, drug or alcohol abuse, current medication and normal orcorrected-to-normal vision. In the first task, participants performed anorthographic judgement task (OJT) requiring a decision about visual sim-ilarity of a pair of characters. In the second task, participants performed ahomophone judgement task (HJT) requiring a decision about the phono-logical similarity of a pair of characters. In the third task, participants per-formed a synonym judgement task (SJT) requiring a decision about themeaning of a pair of characters. Subjects were seated in a comfortablechair in a darkened, sound-attenuated room. EEG was recorded from64 sites using tin electrodes mounted in an elastic cap (NeuroScan) withreferences on the left and right mastoids. The vertical electrooculagram(EOG) was recorded with electrodes above and below the left eye. Inter-electrode impedance was maintained below 5 kX. EEG and EOG wereamplified using a 0.3–40 Hz band-pass and continuously sampled at512 Hz/channel for off-line analysis. Trials with EOG artefacts (meanEOG voltage > ± 100 lV) and artefacts due to amplifier clipping, burstsof electromyographic activity, or peak-to-peak deflection exceeding±100 lV were excluded from averaging.Results and conclusionsTwo ERP components (N1 and N400) were quantified as peak ampli-tude and latency and subjected to repeated measures ANOVA. There wasan effect of hemisphere on the amplitude of N1 F(1,30) = 16.73, p .05). Post hoc analyses using LSD test for multiple compar-ison showed a significant difference between tasks on the amplitude ofN400 F(1,30) = 62.92, p < 0.001 but not N1. Comparisons showed signif-icantly lower amplitude and significantly later N400 for semantic thandoi:10.1016/j.bandl.2006.06.109