What is trimming?
Trimming is Soma’s automated method for excluding excessively slow responses from cognitive task data. These responses often reflect disengagement, fatigue, or attempts to manipulate performance metrics, and do not represent valid cognitive output.
Why is trimming important?
Trimming protects the integrity of your data. It ensures that your athlete’s results are
Accurate — reflecting only real-time, focused responses
Clean — removing outliers that would skew performance trends
Fair — preventing users from slowing down to artificially lower variability or inflate their scores
When does trimming occur?
Certain tasks include a trimming threshold, based on their design. This threshold sets the maximum allowable reaction time — any response slower than this will be excluded. This is not an error — it’s an intentional safeguard to ensure performance data is genuine, consistent, and valid.
Example 1: Minute-on-Minute Trimming
In this graph:
Minute 1 shows a valid RT (~300ms)
Minutes 2 and 3 are marked with red dots — no valid responses were detected
These minutes are excluded from the final data due to excessive delays
Why it matters: Even though the athlete completed the task, Soma detected that their responses were too slow to be meaningful and removed those minutes automatically.
Example 2: Mean Data Trimming
In this session:
Reaction Time = 0 ms
Accuracy = 0%
Speed = 0.00
RCS = 0.00
Lapses = 11
Duration = 3 minute
Interpretation: All responses were trimmed for being too slow.
Trimming Is Not an Error — It’s a Safeguard
When you see 0% accuracy, 0 RT, or red flags in the data, this does not indicate a system bug. It means Soma has filtered out responses that failed to meet the task’s performance threshold.