“…Therefore, extracting the expression for the mean cell delay from (23), substituting and equating the resulting expression to 1 we can solve for (24) Now, we comment on the condition . Since , referring to (22) this condition means the different delay cells are so designed that at time equal they all have the noisevariance parameter as the case when . Finally, if from (23), we extract the expression for the delay cell variance, evaluate the derivative explicitly, substitute , , and (24) in the resulting expression, and normalize by dividing by the mean period, , we obtain the normalized oscillator variance, denoted as (25) where when , the case when we have a constant drift.…”