Individualized Behavioral and Image Analysis of Response Time, Accuracy, and Social Cognitive Load During Social Judgments in Adolescents

DISCUSSION

The results of this study supported the relationship between accuracy, reaction time and question type when analyzing behavioral responses of adolescents when making social decisions. Specifically, the analysis reveled that participants made correct social judgments faster then they made incorrect social judgments and that reaction time was significantly related to question type, with questions with lower social cognitive load taking less time than more difficult social cognitive questions.

The results of this study showed that the number of correct answers decreased over time. These decreasing correct answers could represent the effects of increasing the social cognitive load. This result may have been impacted by the loss of answers from questions of type three when question number 15 from each block from the study was lost because of a software problem with MATLAB. The results also could have been affected by missing data from all stimuli in block two from one the participants. Additionally, results may have also been skewed slightly by the difference in the number of correct versus incorrect answers. There were significantly more correct responses then there were incorrect responses. However, this result is not surprising, as this study employed a test of normal social interaction, and called upon a sample of participants with no history of social problems or deficits.

Correct responses took significantly less time to answer then incorrect responses. It is very interesting that note only did participants take a longer to answer incorrectly, but their increase in amount of time does not appear to have improved their accuracy. There are many reasons why this could have occurred. The first is that the participants either knew the answer, or they did not, right away. When they knew the answer immediately, they did not have to put any thought into generating their response. When they did not know the answer, they may have had to slowly work through the question and wound up getting it wrong regardless of their increasing in time spent considering their response. Inaccurate answers could have also taken longer because the participants were thinking about them harder. Additional analysis of the situation could lead them to both longer reaction times and incorrect answers.

Theoretically, it is not surprising that questions of type one took the shortest amount of time to answer. However, it was a surprising finding that questions of type two and type three took almost the same amount of time. The third question was hypothesized to be the most difficult in the set and required the greatest amount of cognitive load, therefore would take the longest amount of time. This was not the case. It could be that the third question was not as difficult as the researchers intended. It could also be that eventually all difficult social questions reach an asymptote and wind up taking similar amounts of time. Even though questions of type one took the least amount of time, the first question of type one in each block took the same amount of time, or longer, then questions of type two or three. It has been hypothesized that, due to the time break in-between the blocks and the practice sessions that occur between them, restarting the block means that the participants need to readjust their internal timing to the paradigm, and that results in participants taking longer to answer the first question in each block. By the second question in each block, the participants seem to readjust to making the required social judgments.

The results of the behavioral data analysis were used to analyze the functional imaging data of the most accurate and the most innacurate participants. This portion of the analysis was conducted to identify differences in patterns of brain activation when individual response patterns were used as the foundation of the analysis as opposed to a group mean reaction time. As can be seen in the data, the use of individualized response times in the imaging analysis did alter the results of the functional imaging analysis. It should be noted at this point that for the original analysis, summing the data may have produced the best results. Choosing this average decision was helpful in that is allowed researchers to average the data and to look at it as a whole, instead of looking through each individual participant. This current research demonstrates that although a standard response time can be applied as a framework to conduct the functional imaging analysis, it does not take into account the individual variation in performance, both in terms of accuracy and response time, which is important to ensure valid interpretation of imaging findings. Taken together, these results have important implications for modifications in the functional imaging paradigm used to study social cognition using Ciccia's (2006) current protocol.

It is important to note that there were many limitations in this study, including missing data for twenty-four responses to questions of type three due to problems with the response gloves. Because of the missing data, any trends that might have existed specifically for questions of type three may have been skewed. In addition to this, the responses of a singular participant to question block two had to be removed from the study due to inappropriate responses (i.e. failing to answer a question or answering both 'yes' and 'no'). Although the information with the VR gloves was accurate enough for this study, the next study will change to a button response box. The researchers believe that this will be more successful within this study for a number of reasons. The first is that a button-pressing response paradigm is more natural, and might require less training and concentration then just moving a finger. The use of the response box could lead to fewer instances of undecided (answering both yes and no) answers. The response box also allows for more rigorous accounting of reaction time. Unlike the present system, where a movement of a gloved hand causes a waveform pattern which indicates the moment of response on a continuous time scale, the button system would log the exact time of the response (reducing the time required to go to the waveform image and manually note the response time). This would give a more accurate account of response time, and, since the system is disabled after a participant responds to a question set, there would be fewer instances of undecided responses.

Although the methodology of the original study worked well, this research demonstrates that more specificity is needed when it comes to social cognitive functional imaging analysis. The next step in this study will be to do a full analysis of the imaging data taking individual reaction time and accuracy into account and using the individual analysis to conduct a group data analysis. In addition, the improvements in the current paradigm will allow for application of this paradigm with an adolescent clinical population that has know social cognitive deficits, such as Autism or Asperger's syndrome.

The results of the present study will help shape continued work in this lab in the areas of methodological improvement, in-depth comparative behavioral and imaging data analysis, consideration of developmental effects on neural activation, and applications to a variety of clinical populations with social cognitive deficits. Specifically, the lab hopes to due an individual-based functional imaging analysis on the effects of accuracy on individuals region of interest activation. The lab also plans to study inactivation in the group data as well as in the individual data.

Acknowledgements

I would like to thank Dr. Ciccia, Dr. Dzemidzic, Dr. Rothenberg, Dr. Wiliams, Dr. Tkach, and The Communication Sciences Department at Case Western Reserve University, for all of their help on this project. This work was supported by the American SpeechLanguage Hearing Foundation New Investigator's grant to the Second author and by a Howard Hughes Medical Institute (HHMI) Fellowship awarded to the first author.

References

1. Abu-Akel, A. (2003). A Neurobiological Mapping of Theory of Mind. Brain Research Reviews 43: 29-40.
2. Adolphs, R. (1999). Social Cognition and the Human Brain. Trends in Cognitive Science 3(12): 469-479.
3. Amodio, D. M., Frith, C.D. (2006). Meeting of Minds: The Medial Frontal Cortex and Social Cognition. Nature Reviews Neuroscience 7: 268-277.
4. Basacki, S., & Astington, J. W. (1999). Theory of mind in preadolescence: Relations between social understanding and social competence. Social Development 8(2): 237-255.
5. Ciccia, A., Dzemidzic, M., Lowe, M., Horenstein, H. (2006). Neural substrates of increasing social cognitive demand in older adolescents. Conference Proceedings from Human Brain Mapping, Florence, Italy.
6. Frith, C., & Frith, U. (1999). Interacting minds A biological basis. Science 286: 1692-1695.
7. Frackowiak, C., Friston, K., Frith, C., Dolan, R., Price, C., Zeki, S., Ashburner, J., & Penny, W. (2004). Human Brain Function, 2nd Edition. Elsevier, London, UK.
8. Leslie, A. M. (2005). Developmental parallels in understanding minds and bodies. Trends in Cognitive Science 9(10): 459-462.
9. Saxe, R. (2006). Why and How to Study Theory of Mind. Brain Research 1079: 57-65.
10. Thesen, S., Heid, O., Mueller, E., & Schad, L. (2000). "Prospective acquisition correction for head motion with image-based tracking for real-time fMRI" Magnetic Imaging Resonance Medicine 44: 457-465.
11. Turkstra (2000). Should my shirt be tucked in or left out? The communication context of adolescence. Aphasiology 14(4): 349364.
12. Turkstra, L., Ciccia, A., Seaton, C. (2001). Interactive behaviors in adolescent conversation dyads. Language, Speech, and Hearing Services in Schools 34, 117-124.