How Do You Detect Confusion in Code Reviews?

There is a theoretical model of Michelle Jordan which defines it as situations when individuals have a sense of wandering, doubt or unease about how to will unfold. So this was essentially the theoretical model we have used and based on this model, we have mixed automatic detection of certain keywords with manual labeling in code reviews. And so based on these kinds of questions, we have obtained more examples of confusing situations than we would have been able to detect directly by observing code review comments.