Resumen de Procesamiento léxico-semántico en un grupo de sujetos sanos: estudio con potenciales relacionados a eventos
Esther Baldera Cruz, Josefina Ricardo Garcell, Gerardo Heinze, Gabriela Galindo Villa Molina
- español
Introducción:
El conocimiento del significado de las palabras es uno de los procesos centrales de la memoria semántica. Para evaluar el acceso a la representación cognoscitiva del significado de las palabras, en el presente estudio se utilizó el paradigma de decisión léxica desarrollado por Marcos, donde el sujeto debe reconocer si el estímulo presentado corresponde a una palabra o a una pseudopalabra. Lo anterior se hizo con el propósito de contar con un modelo de procesamiento normal que sirviera para contrastar los hallazgos con patologías en que se encuentra alterada la memoria semántica.
Método:
La muestra estudiada estuvo conformada por 32 sujetos sanos (7 hombres y 25 mujeres), diestros y sin antecedentes personales o familiares de padecimientos neurológicos o psiquiátricos. La edad promedio fue de 34.4 (+ 9.56) años y el promedio de escolaridad fue de 16.2 (+ 4.4) años.
El paradigma de decisión léxica empleado está constituido por 408 estímulos: 240 palabras y 168 pseudopalabras. Los criterios para seleccionar las palabras fueron frecuencia, longitud, catego ría gramatical y morfología.
Se obtuvo el registro monopolar del electroencefalograma (EEG) a partir de 19 derivaciones (F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz y Pz) y los potenciales relacionados a eventos (PREs) para los subestados: palabra y pseudopalabra.
Resultados:
Los datos conductuales indicaron que la escolaridad guarda relación con el reconocimiento de las palabras y que éstas se reconocen más rápido que las pseudopalabras.
Para determinar las diferencias entre los PREs promedios de ambos subestados, se aplicó la prueba t de Student con la corrección de Bonferroni, con un nivel de significancia de p < 0.0002. Se observaron diferencias significativas entre los dos subestados y este resultado no se vio afectado por la edad ni por el sexo de los sujetos.
En el intervalo de entre 375 y 495 ms de latencia se aprecia un componente negativo ante la presencia de las pseudopalabras, que muestra diferencias significativas en las derivaciones F3, F4, C3, C4, P3, P4, O1, O2, T3, T5, T6, Fz, Cz y Pz. Las diferencias de amplitud entre los dos subestados fueron más evidentes en las derivaciones Pz y P3, seguidas por Cz.
También se observó un componente positivo en el intervalo de entre 700 y 795 ms de latencia ante la presencia de las pseudopalabras, donde las diferencias significativas se aprecian en las derivaciones F3, C3, P3, F7, T3, T5 y Pz. Las diferencias de amplitud entre los dos subestados fueron más evidentes en las derivaciones Pz y P3, seguidas por C3.
Discusión:
Al analizar los aspectos conductuales, se observó que los sujetos cometieron más errores en el caso de las palabras. Sin embargo, se equivocaron más los de menor escolaridad, de lo que se infiere que esta variable se puede relacionar con la extensión del repertorio léxico. Asimismo, se encontraron diferencias significativas en el tiempo de reacción, ya que las palabras y las respuestas correctas propician tiempos de reacción más cortos.
Por otro lado, debido a que tanto el componente negativo como el positivo aparecieron mucho antes que la respuesta, se considera que la actividad obtenida por los PREs se relaciona con el procesamiento cognoscitivo de la información frente al paradigma utilizado.
De acuerdo con los datos de investigaciones previas, el componente negativo parece estar relacionado con la respuesta generalizada de la actividad cerebral ante un estímulo carente de significado, es decir, de la pseudopalabra. Asimismo, parece que la amplitud de la onda se relaciona con la cantidad de activación necesaria para acceder a la representación semántica del estimulo en la memoria.
Con respecto al componente positivo en esta investigación, éste puede relacionarse con los recursos atencionales necesarios para procesar la aparición de las pseudopalabras, por lo que se interpreta como una P300 tardía (P3b).
- English
Introduction:
An insight into the meaning of words is one of the central processes of semantic memory. To evalúate the access to the cognitive representation of the meaning of words, in the present study we used the lexical decision paradigm developed by Marcos. In this situation, the subject has to recognize if the presented stimulus corresponds to a word or a pseudo-word with the purpose of building a model of normal processing. Once such a model of normal processing is obtained, the findings can be contrasted with pathologies in which semantic memory is altered.
Method:
The sample consisted of 32 healthy subjects (7 men, 25 women), right-handed and with no personal or familial history of neurological or psychiatric conditions. The average age of the subjects was 34.4 (+ 9.56) years and they had an average educational level of 16.2 (+ 4.4) years.
The lexical decision paradigm employed in this study is constituted by 408 stimuli, 240 words and 168 pseudo-words. The criteria for word selection were: frequency, length, grammatical category and morphology.
Electroencephalogram (EEG) monopolar recording was obtained from 19 derivations (F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz and Pz), as well as event-related potentials (ERPs) for the word and pseudo-word sub-states.
Results:
In the first place, a chi-squared analysis was performed to establish whether significative differences existed between the rates of correct and incorrect answers for both sub-states. The value of chi-squared was 65.7 (gl=1) and significant for p<0.0001.
A correlation value of 0.43 (p< 0.02) was found when the educational level and the percentage of correct answers in the sub-sate word were compared. On the other hand, for the pseudo-word sub-state, the value 0.24 was encountered for the same correlation, being statistically non-significant.
Pearson's correlation coefficient was also calculated for the educational level variable compared to the mistakes committed when subjects were presented with frequent and infrequent words.
In the case of infrequent words, a value of r = - 0.43 (p<0.02) was obtained when the educational level and the number of mistakes were correlated. No correlation was found when the educational level and the number of mistakes commited for frequent words were compared (r = - 0.06).
A multivariate variance analysis for repeated measures was performed to determine significant differences between the reaction times when recognizing words or pseudo-words. The outcome showed that all effects were significant in the following cases: reaction times for words and pseudo-words, notwithstanding whether they were correct or incorrect; comparison between correct and incorrect answers, independently of their being words or pseudo-words, as well as the interactions between both effects.
To determine differences between average ERPs for both sub-states, Student's T-test was applied with Bonferroni's correction and p<0.0002 as the significance level. Significant differences were encountered between the two sub-states, independently of the age or gender.
In the 375-495 ms latency interval, a negative component was appreciated in the pseudo-words case, showing significant differences (p<0.0002) in the following derivations: F3, F4, C3, C4, P3, P4, O1, O2, T3, T5, T6, Fz, Cz and Pz. Amplitude differences between the two sub-states were more evident in Pz and P3 derivations followed by Cz.
In addition, a positive component in the 700-795 ms latency interval was detected (mainly in 795 ms) when pseudo-words were presented. Here, the significant differences (p<0.0002) were manifest in the following derivations: F3, C3, P3, F7, T3, T5 and Pz. Amplitude differences between the two sub-states were mainly patent in Pz and P3 followed by C3.
Discussion:
When analyzing behavioral aspects, subjects made more mistakes when presented with words. However, individuals with less education were the ones committing more mistakes. From this we can infer that this variable may be associated with the range of lexical repertoire.
A relation was encountered between educational level and word recognition. With regard to reaction times, significant differences were detected between both sub-states, since the recognition of both words and correct answers was achieved in shorter reaction times.
Average reaction times for words and pseudo-words were 819.73 ms and 999.35 ms, respectively. Similarly, the latest potential component appeared in an interval of significant differences between 600 and 940 ms, though with a significance p<0.0005 between 690 and 805 ms. This means that positiveness occurred much sooner than the response, implying that the activity underlying ERPs is related to a cognitive processing of information due to the paradigm used.
The analysis of ERPs primary components for both sub-states shows that significant differences arise until 270 ms.
The negative component in this study was present between 270 and 580 ms, rendering it similar to N400 given its latency (around 400 ms). Although well-defined in centro-parietal regions, its distribution was generalized, which corresponds to the results of studies using the semantic incongruence paradigm.
According to the data from previous research on ERPs, N400 has been associated with the integration process. If this were the case, this association would be equivalent to the semantic incongruence within a lexical integration process described in conventional literature as a "semantic facilitator", only that this time it would be limited to the process of access to the lexicon, which can be interpreted as a discrimination of the answer by assigning a meaning to a word, that is, to process information in the semantic module.
This negative component may be related to the generalized response to brain activity when given a meaningless stimulus, i.e., a pseudo-word. Similarly, the wave amplitude may be related to the amount of activation necessary to gain access to the semantic representation of the stimulus in the memory.
With regard to the positive component in this study, it is present between 600 and 940 ms and is interpreted as a late P300 (P3b), which has a latency in the 500-1400 ms interval. It is distributed over the centro-parietal region, making it a liable participant in the task categorization process, in which it is necessary to discriminate between the target from the non-target stimulus, and also reflects focalized attentional processes (voluntary) involved in the execution of the task.
From the former, it is believed that this component may be related to attentional resources necessary to process the presentation of pseudo-words.
Research dealing with the P600 component locate it within the context of statements and associate it with an anomaly in statement syntax. Therefore, even though the positive component lies within the P600 latency domain, this particular component was not considered as being present in this study, because a syntax incongruence paradigm was not used.
Finally, the contribution of the present study lies in the finding of N400 and "P600" components, which have been reported when the "semantic facilitator" and the syntax incongruence paradigm were respectively used, but had not been observed when a lexical decision paradigm based on word recognition per se was utilized. Similarly, given that our results stem from a sample of healthy subjects, a comparison can be made with a patient population with semantic memory alterations.
