Presented By: Cognition & Cognitive Neuroscience
CANCELLED: CCN Forum:
Tessa Abagis and Lauren Grant, CCN Graduate Students
Note: This event has been cancelled.
Ed Smith Neuroscience Award Talks
Tessa Abagis
Title:
Investigating external and internal distraction in adults with ADHD
Abstract:
Attention deficit/hyperactivity disorder (ADHD) is often considered to be a disorder in children and adolescents but is in fact diagnosed in 2.5% of adults. Adults with ADHD experience poorer resistance to external task-irrelevant distraction (Forster and Lavie, 2016) as well as to internal task-irrelevant distraction, or mind-wandering (Franklin et al., 2017). Past work in the lab has determined a visual search task that displays robust performance differences between adults with and without ADHD due to singleton distractors. In the work I will discuss, I have employed both eye-tracking and mind-wandering methods during this visual search task to evaluate how external and internal distraction affect task performance in adults with and without ADHD.
Lauren Grant
Title:
Outliers Among Us: How to Identify and Deal with Extreme Data Points in ECoG
Abstract:
Outliers in experimental data can result in false positives (Type I errors) or false negatives (Type II errors) that dramatically change a researcher’s conclusions. Given the adverse effects of publishing such distorted conclusions (e.g., contributing to the replication crisis), many researchers try to identify and remove outliers before conducting their main statistical analyses. However, there are a large number of outlier removal methods in the literature, and it remains unclear which ones are most effective. Along these lines, my recent work with reaction time (RT) data suggests that employing Sn – a robust estimator of scale – minimizes both Type I and Type II errors in subsequent data analyses, relative to several alternative methods. However, it is unclear whether Sn outperforms competing outlier removal methods for other data types. To test this hypothesis, I conducted a series of simulations in the context of a typical ECoG experiment. Interestingly, the median absolute deviation (MAD) – rather than Sn – was a highly effective method for dealing with extreme data points. Such findings reveal that there is no “one shoe fits all” solution to the problem of removing outliers. Rather, the most effective method varies with the nature of the data itself.
Ed Smith Neuroscience Award Talks
Tessa Abagis
Title:
Investigating external and internal distraction in adults with ADHD
Abstract:
Attention deficit/hyperactivity disorder (ADHD) is often considered to be a disorder in children and adolescents but is in fact diagnosed in 2.5% of adults. Adults with ADHD experience poorer resistance to external task-irrelevant distraction (Forster and Lavie, 2016) as well as to internal task-irrelevant distraction, or mind-wandering (Franklin et al., 2017). Past work in the lab has determined a visual search task that displays robust performance differences between adults with and without ADHD due to singleton distractors. In the work I will discuss, I have employed both eye-tracking and mind-wandering methods during this visual search task to evaluate how external and internal distraction affect task performance in adults with and without ADHD.
Lauren Grant
Title:
Outliers Among Us: How to Identify and Deal with Extreme Data Points in ECoG
Abstract:
Outliers in experimental data can result in false positives (Type I errors) or false negatives (Type II errors) that dramatically change a researcher’s conclusions. Given the adverse effects of publishing such distorted conclusions (e.g., contributing to the replication crisis), many researchers try to identify and remove outliers before conducting their main statistical analyses. However, there are a large number of outlier removal methods in the literature, and it remains unclear which ones are most effective. Along these lines, my recent work with reaction time (RT) data suggests that employing Sn – a robust estimator of scale – minimizes both Type I and Type II errors in subsequent data analyses, relative to several alternative methods. However, it is unclear whether Sn outperforms competing outlier removal methods for other data types. To test this hypothesis, I conducted a series of simulations in the context of a typical ECoG experiment. Interestingly, the median absolute deviation (MAD) – rather than Sn – was a highly effective method for dealing with extreme data points. Such findings reveal that there is no “one shoe fits all” solution to the problem of removing outliers. Rather, the most effective method varies with the nature of the data itself.
Co-Sponsored By
Explore Similar Events
-
Loading Similar Events...
