 |
 |
Bob, I would suggest that you model the ISI duration in the ISI events. The reason is that an impulse will not capture the ISI time period very well. You could then drop #2 (Total Duration). If you look at the HRF of a impulse compared to 5.5seconds, you will notice significant difference. As the responses are variable durations, I'm not sure collinearity will be a big issue. Best Regards, Donald McLaren ================= D.G. McLaren, Ph.D. Research Fellow, Department of Neurology, Massachusetts General Hospital and Harvard Medical School Postdoctoral Research Fellow, GRECC, Bedford VA Website: http://www.martinos.org/~mclaren Office: (773) 406-2464 ===================== This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (773) 406-2464 or email. On Mon, Feb 4, 2013 at 12:05 PM, Bob Spunt <[log in to unmask]> wrote: > Dear experts, > > I am writing to see if anyone can share their thoughts on my analytical > approach. In an event-related design, participants made manual responses to > three stimulus types: A, B, and C (stimulus onset asynchrony varied from 6.5 > to 9.5 s, mean of 7.5 s). At response, stimuli were replaced by a fixation > cross until the onset of the next stimulus. My research question regards > these interstimulus intervals (ISIs). Namely, is variation in BOLD activity > before stimulus onset (i.e., during the ISI) associated with response time > (RT) to stimulus onset? I want to estimate this association for each of the > three stimulus types. > > I am currently using a parametric modulation analysis, with three regressors > modeling the response to each of the three stimulus types (variable epoch, > duration = RT) and three regressors modeling the response to the ISI > preceding each of the three stimulus types (impulse function). (I should > briefly note that trials with outlier RTs are removed and modeled in a junk > regressor.) I chose to model the response-to-ISI as an impulse function to > minimize multicollinearity in the design matrix. Finally, I included three > additional regressors modeling variability in the response to the ISI as a > function of the following parameters (all de-meaned): > > 1. RT to the next stimulus (the regressor of interest) > 2. Total duration of the ISI > 3. A binary variable indexing whether or not the prior stimulus was of the > same type as the next stimulus > > With this approach, the models estimate and I am actually getting reasonable > results. However, given that I have not found any precedent for this > approach in the literature, I am writing to see what others think. If you > have any thoughts, I would be grateful to hear them. > > Cheers, > Bob >