tag:blogger.com,1999:blog-5427587583898590293.post2573885714930787413..comments2023-10-31T03:41:34.562-07:00Comments on Brad Buchsbaum's Blog: Reaction Time Experiments: Functional Neuroimaging on the CheapBrad Buchsbaumhttp://www.blogger.com/profile/10757537675625801119noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-5427587583898590293.post-50748425025897652102013-04-11T08:49:36.833-07:002013-04-11T08:49:36.833-07:00If they answer with two fingers, doesn't can&#...If they answer with two fingers, doesn't can't we just say that counts as different hemispheres, and thus two voxels?Anonymoushttps://www.blogger.com/profile/11147404084121949626noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-17340920081689574522013-04-06T17:04:13.978-07:002013-04-06T17:04:13.978-07:00According to Brad “The logic of an RT experiment i...According to Brad “The logic of an RT experiment is simple. When a subject responds more slowly in one condition than another, we say that this condition required more ‘cognitive processing’.“ This is not correct. No cognitive psychologist would say this. The idea that cognitive processing is some monolithic ability and that the only way cognitive tasks differ from each other is in the quantity of this ability they need is entirely foreign to cognitive psychology. The reason is obvious: if I say “I can explain why RTs are slower in condition A than condition B It is because condition A requires more cognitive processing”, you might then ask “How do you know that condition A requires more cognitive processing than condition B?” The only answer I could give – that the evidence that this is so is that RTs are slower in condition A – is of course embarrassingly circular.<br /><br />What cognitive psychologists actually use RT data for is to draw inferences about the functional architecture of whatever cognitive system subjects are using to perform the experimental task. They use RT data to try to answer such questions as: is retrieval from STM a parallel or a serial process? Is this retrieval self-terminating or exhaustive? What are the roles of featural and holistic processing in face recognition? Is spoken word recognition a purely feedforward process or is feedback involved? And so on. Different theoretical answers to such questions make different predictions about how RTs will differ between conditions, and so the presence or absence of such differences can be used to adjudicate between competing theories. That’s what the logic of RT experiments is.<br /><br />Only when this correct characterization of what cognitive psychologists actually do is adopted can one then ask the question: has it yet happened that data from functional neuroimaging has successfully been used to adjudicate between competing theories, given that RT data has often been used successfully for this purpose? Brad quotes me as claiming that “functional neuroimaging has nothing to say when it comes to deciding between two psychological theories”, but that’s not quite right either: it is missing the crucial word “had”. What I have argued in various places is that “functional neuroimaging has had nothing to say when it comes to deciding between two psychological theories”.<br />Max Colthearthttp://www.maccs.mq.edu.au/~max/noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-5128170113693501782013-03-22T21:05:10.378-07:002013-03-22T21:05:10.378-07:00Sadly Brad, RTs only correlate weakly with neurona...Sadly Brad, RTs only correlate weakly with neuronal spiking, so Mike PAge will have none of it...<br /><br />http://bit.ly/10v0XARAnonymoushttps://www.blogger.com/profile/12165546420696517053noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-35221620477756871212013-03-22T09:32:54.947-07:002013-03-22T09:32:54.947-07:00it's all about branding....it's all about branding....Brad Buchsbaumhttps://www.blogger.com/profile/10757537675625801119noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-45271736433289194772013-03-22T09:31:33.370-07:002013-03-22T09:31:33.370-07:00The RT-imaging community prefers to keep it simple...The RT-imaging community prefers to keep it simple: one voxel. Still, there is a compelling case for multi-finger pattern analysis (MFPA), as it has been claimed to increase sensitivity, and offers the option of presenting results using a heatmap, which is nice. Brad Buchsbaumhttps://www.blogger.com/profile/10757537675625801119noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-19171756071289614052013-03-22T09:20:59.914-07:002013-03-22T09:20:59.914-07:00pencil-and-paper survey = graphite-modulated refle...pencil-and-paper survey = graphite-modulated reflected-light imaging (GMRLi) <a rel="nofollow">#neurofriedpsychology</a>Anonymoushttps://www.blogger.com/profile/16433932891639112214noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-61320539359778422312013-03-22T08:45:57.968-07:002013-03-22T08:45:57.968-07:00aren't there as many voxels as there are effec...aren't there as many voxels as there are effectors? will we start having to do fingerwise error rate correction?Russ Poldrackhttps://www.blogger.com/profile/03305657400941743430noreply@blogger.comtag:blogger.com,1999:blog-5427587583898590293.post-51405678629106815242013-03-22T07:47:38.600-07:002013-03-22T07:47:38.600-07:00I can't wait to submit my next fRTI (functiona...I can't wait to submit my next fRTI (functional reaction-time imaging) paper to Nature Neuroscience!<br /><br />BTW let us not forget about other "cheap neuroimaging" techniques with better spatial/temporal resolution than fRTI, e.g., such as ERO (event-related optometry, i.e., eyetracking).Anonymoushttps://www.blogger.com/profile/16433932891639112214noreply@blogger.com