| Background:The mental health of military personnel has always been a key area of concern for the military,and screening and attention to key populations is an important part of both psychological selection and the development of psychological work.Key populations are those with psychological problems and mental disorders,which can have a significant negative impact when they enter the military or when they are neglected.People with impulsive personality tendencies are included in this study because they are associated with suicidal tendencies and various psychological disorders,and their impulsive behaviors are also risky and therefore belong to the priority population.Among the current screening instruments for people with impulsive personality tendencies,the self-report method,i.e.,the self-report scale,is the most widely used and easy to implement,but has shortcomings such as social approbability and self-masking.As a refinement of the self-report method,the use of a linguistic material that elicits cognition and emotion from the test taker may improve the scale to some extent and thus complement it.In this study,we used impulsive personality dispositions as the target population,and established a special emotional situational language material in a military setting through a qualitative study,aiming to induce impulsivity in impulsive personality dispositions by presenting emotional situational language material.psychometric accuracy and validity.Subjects and Methods:The study population consisted of both impulsive and non-impulsive individuals,and the study was divided into three parts.The first part was the construction of emotional situational language materials.Using a qualitative research method,impulsive personality predispositions in the military were interviewed to understand their impulsive emotions,behaviors,or cognitions in specific emotional situations,and after machine coding and coding optimization,the emotional situational language materials were compiled.The second part is an analysis of the differences in the eye-movement characteristics between impulsive personality tendencies and non-impulsive individuals when reading the emotional situational language materials and the scale language materials.In addition to the analysis of the response indicators of the evoked effect of the situational language materials,the eye movement indicators of the two groups were collected objectively through the eye-tracking technique,and the indicators were analyzed to explore the differences in subjective perceptions and eye movements of the two groups when reading the language materials,and to explain the impulse-evoking effect of the language materials on them.The third part compares the performance of the classifier model for the emotional-situational language materials with that of the scale language materials.The classifier models were built by machine learning of the eye movement indicators obtained in the second part,and the classifier performance was compared to verify the impulse-evoking effect of the emotional-situational language materials on impulsive personality tendencies.Results:Part I: People with impulsive personality tendencies in the military produce impulsive emotions,behaviors,or cognitions in response to particular situational stimuli,and the emotional situational verbal material was initially assessed well(all p<0.01)Part II: Emotional situational verbal materials had better impulse elicitation effects than scale verbal materials for military impulsive personality tendencies for.(1)There was a significant difference in the scores of impulsive personality predisposed individuals and non-impulsive individuals with respect to their scores induced by emotional situational language materials(t = 20.95,p < 0.001).(2)Mean eye-movement gaze duration under emotional-situational verbal material stimulation was significantly interacted by group and interest area [F(1,339)= 7.522,p =0.006,ηp2 = 0.022],with statistically significant differences in simple effects between the two groups within the stem interest area(t = 1.973,p < 0.05)and within the option interest area(t = 5.655,p < 0.01).(3)Eye-movement gaze duration with emotional-situational verbal material stimuli,group interaction with interest area was significant [F(1,339)= 38.758,p < 0.001,ηp2 =0.103],and the difference in the simple effect within the stem interest area was statistically significant between the two groups(t = 9.468,p < 0.01).(4)Eye-movement gaze return time with emotional-situational verbal material stimuli,group interaction with interest area was significant [F(1,339)= 173.708,p < 0.001,ηp2 =0.339],and the difference in the simple effect within the stem interest area between the two groups was statistically significant(t = 19.283,p < 0.01).(5)Pupil change values in response to emotional situational verbal material stimuli,with a significant group-area of interest interaction [F(1,339)= 15.416,p < 0.001,ηp2 =0.043],with statistically significant differences between the two populations for simple effects within the stem area of interest(t = 9.442,p < 0.01)and within the option area of interest(t = 4.894,p < 0.01).(6)Mean eye movement gaze duration with scale verbal material stimuli,group interaction with interest area was significant [F(1,339)= 252.251,p < 0.001,ηp2 = 0.427],and the difference in simple effects within the stem interest area was statistically significant between the two groups(t = 2.327,p < 0.05).(7)The interaction between group and interest area was significant for oculomotor gaze time with scale verbal material stimulation [F(1,339)= 28.806,p < 0.001,ηp2 =0.078],and the difference in the simple effect within the stem interest area was statistically significant between the two groups(t = 8.343,p < 0.01).(8)Scale verbal material stimulation with eye movement gaze back time,group interaction with interest area was significant [F(1,339)= 15.293,p < 0.001,ηp2 = 0.043],and the difference in the simple effect within the stem interest area between the two groups was statistically significant(t = 5.288,p < 0.01).Part III: Classifier models built for emotionally situational linguistic material had higher recognition rates for.(1)The highest recognition accuracy of the emotional-situational language material was 0.912 on the four models of decision tree,random forest,KNN,and support vector machine;the highest recognition accuracy of the scale language material was 0.573.(2)The highest AUC value of emotional-situational language materials under the ROC curves established by the four models of decision tree,random forest,KNN,and support vector machine was 0.968;the highest value of scale language materials was0.580.The difference of AUC values between the two types of language materials was statistically significant(all p<0.001).Conclusion:1.special emotional situational language materials showed stronger impulse-evoking effects on military impulsive personality tendencies compared to scale language materials on both eye movement and response indicators,with significant discriminatory effects on impulsives and non-impulsives.2.The classifier model established by machine learning of the eye-movement indicators fond that the emotional situational language materials have better recognition accuracy for impulsive personality tendencies than the scale language materials,and canachieve a certain degree of recognition of impulsive personality tendencies by eye-movement characteristics(the highest recognition rate was 0.912). |
PDF Full Text Request