====== PSYC 410 fMRI Project ====== You are all no longer Padawan learners, but have become young fMRI Jedis (that's right, I'm using Star Wars references. I apologize for nothing!!). In this last project you will show off your newly acquired mastery of the brain imaging force (okay okay, I'll stop now) and perform a complete fMRI analysis. This fMRI report is due on __Wed Apr 22nd @ 7:00 pm__ See [[#Turning in your assignment|below]] for specific instructions on how to submit. This assignment will be very similar to the lab reports you've done throughout the semester. The primary difference is that I will not furnish you with a step-by-step guide for carrying out the analyses. For the fMRI analysis you will act as if you're newly minted fMRI researchers, not students. However, there is no scientist who knows and remembers everything and can work in total isolation without the aid of colleagues, references, and the internet. So you are allowed to use any resource you'd like, including our lab wikis, to assist you in your analyses. This also means you are permitted to ask each other for assistance. However, you are not permitted to work together. That is, asking for a little help here and there when you get really stuck with something is fine, but I expect you to primarily work independently. **What do I mean when I say you can each other for a "little help", but that you are not permitted to work together?** Pretend your classmates are professional colleagues with whom you have a friendly relationship. You would definitely not hesitate to reach out to such a colleague if you hit a brick wall and needed some advice. But you certainly would //not// fill up their inbox and expect them to hold your hand through your analysis. Go Slow. Be Methodical. Go Slow. Be Methodical. I really can't stress this enough. You have all experienced first hand how the seemingly smallest mistakes can cause fMRI analysis to go off the rails. What's worse is that sometimes this happens in ways that aren't so obvious. Your analysis might complete without errors, but you'll end up with invalid results. Go Slow. Be Methodical. Go Slow. Be Methodical. Setting up analyses can be repetitive and tedious, so it's only natural to want to breeze through it. But you must fight the temptation to just click-click-click-Go. I can't begin to count the many dozens of hours I've spent //re-doing// analyses because I was hasty and careless the first time. Think about each setting as you set it. Make sure you understand **why** you're using it, and that you're setting it correctly. Go Slow. Be Methodical. ===== Turning in your assignment ===== * //fMRI Report//: Turn in your [[#fmri_report|report]] via Moodle * //fMRI Analysis//: see below * Create a Google Drive folder and share it with me. * Name the folder ''yourlastname_410_fmri'' * Create three sub-folders: * ''first-level'' * ''second-level'' * ''third-level'' * //First-level analysis//: Put your design file (.fsf file) for each run in the ''first-level'' folder. * //Second-level analysis//: Put your design file (.fsf file) in the ''second-level'' folder. * //Third-level analysis//: Put your design file (.fsf file) //**and**// a zipped version of the ''cope3.gfeat'' directory in the ''third-level'' folder You must use the follow directory names. Failure to do so will result in a loss of points. When I evaluate your analysis I will write programs to assist me. These programs will assume that the data are in the directories exactly as specified below. If you use a different name you will not get points for that section. For instance, if you enter ''run2'', instead of ''run02'' you will lose points. Throughout the wiki, replace: * **''SUBJ''** with the ID of your assigned subject * **''TASK''** with the task name (''motion'') * **''RUN''** with whatever run you are analyzing (e.g., ''run02'') - Create a subject directory on your computer. mkdir -p ~/Desktop/yourLastName_fmri_proj/SUBJ - Create a third-level output directory on your computer. mkdir -p ~/Desktop/yourLastName_fmri_proj/third_level /* SHOULDN'T BE NECESSARY ==== USB Drive error ==== ISSUES WITH USB DRIVE If you are having issues with your USB drive such that it has appended a '' 1'' to the drive name (e.g., ''lil_andrew 1'', then try the following fixes. **FIX 1** * Unmount the drive by dragging it to the recycle bin * Physically remove the drive * Restart computer * If you're unable to restart without an admin password, then just log out and log back in * Re-insert the drive Hopefully this will fix the problem and remove the '' 1''. It's possible that your drive name will appear in a ll uppercase (e.g., ''LIL_ANDREW''), but this should be okay. //When you are specifying your output directory, still specify using all lower case!// **FIX 2** Fix 1 should work. If for some reason it does not ... * Create a temporary directory on your desktop and move the entire contents of your USB drive to that directory * In your terminal, enter ''open /Applications/Utilities/Disk\ Utility.app'' * In the left column highlight your drive (see image below). * Click the ''Erase'' button * Change **Format** to ''Mac OS Extended (Case-sensitive, Journaled'' * Put in the correct drive name (e.g., ''lil_andrew'') * Click ''Erase'' * That should do it. However, sometimes it might fail and say that it was unable to erase the drive or some sucvh thing. If os, just repeat these same steps a second time. Now it will magically work. Computers, amiright!? * **Move the files you put into the temporary directory on your desktop back onto the USB drive** Of course, if you are still having an issue you should email me ASAP. {{ :psyc410_s18:fix1.jpeg?400 |}} {{ :psyc410_s18:fix2.jpeg?400 |}} */ ===== Point-Light Display Paradigm ===== ==== The Gist ==== It has long been known that specific parts of the visual system are particularly sensitive to the perception of motion. More recently it has been discovered that the brain also has regions that are specifically sensitive to biological motion (see [[#Example Videos|example]] below). This is analogous to face perception in that we know the visual system has some areas that prefer objects to non-objects, but that face "objects" have their own dedicated region. Point-light display movies are used to identify these biological motion sensitive areas. The movies are simply lights on a black background. Nonetheless, when these lights are placed at particular locations on a body in motion, we easily perceive it as a body and not just randomly moving lights. So why use point-light movies instead of just a video of a person walking? Good question. The reason point-light movies are great is that it is very easy to create an excellent control condition. The only stimulation reaching your eyes are discs of light in motion. To create a great control condition we simply need to scramble around the location of those discs (see example below). Doing so totally disrupts the perception of biological motion, but preserves all other facets of the display (total luminance, number of items, speed and direction of the individual lights, etc.) If you're interested in seeing just how powerful these simple motion cues are, check out [[http://www.biomotionlab.ca/Demos/BMLwalker.html|this site]] on which you can change things like gender and mood! ==== Example Videos ==== Check out examples of what these movies look like: * ''Right click'' on each of the movies below and select ''Loop'' (Use Google Chrome) * Play both movies * The movie on the left shows //biological motion//. It is clearly a man walking from right to left. * The movie on the right shows //non-biological motion//. It has all the same elements of the 'walker' video, but these elements are scrambled, thus disrupting the perception of biological motion. {{:psyc410:movies:walker.mp4}} {{:psyc410:movies:scrambled_walker.mp4}} ==== Common Results ==== Biological motion is known to activate a region of the lateral temporal cortex (particularly in the right hemisphere) that is very near two related areas; the posterior superior temporal sulcus, which activates to faces, and area MT/V5, which is the motion sensitive area of the visual system. {{:psyc410:documents:bonda_1996.pdf|This paper}} is the first to show this using neuroimaging (this was in 1996, so they used Positron Emission Tomography, not fMRI). In this image (Engell and McCarthy, 2013, //NeuroImage//), you can see regions of the brain that have the highest probability of being activated by biological motion more strongly than by non-biological motion. You can see that this set of areas is __very__ similar to the one activated preferentially by faces. Activation in the right hemisphere includes the fusiform gyrus, the posterior superior temporal sulcus, area MT/V5 (which is selective for motion in general), and the "extrastriate body area" (which is preferentially activated by the perception of body parts). {{:psyc410:images:bio_prob_lh.png?200|}} {{:psyc410:images:bio_prob_rh.png?200|}} Another __very cool__ way to see what results you should expect is to check out [[http://neurosynth.org/analyses/terms/biological/]], which is kinda like a Google search engine for brain activations. The link will take you to the activations associated with biological motion. See [[:psyc410:neurosynth|here]] for a primer on Neurosynth. ===== Data Used For This Project ===== ==== Participants ==== /* I don't actually know the participant demographics so I just made this up for their report */ * Seventeen participants * All right-handed * 10 female, 7 male * Mean age = 20.4 years ==== Task Parameters ==== Pay Attention to Task Parameters These parameters may be similar to the block design you've analyzed in the last few weeks, but they are not necessarily identical. Subjects in the this study were instructed to view short movies depicting point-light displays of **//biological// ** or **//non-biological//** (random) motion. Between blocks there were //rest// periods in which the subjects viewed a small fixation cross on a grey background. Occasionally a "movie" would appear in which nothing was moving. Subjects were instructed to count how many of these static movies appeared and report this number at the end of each run. This task is not relevant for our investigation. Rather, it was used solely to ensure the participant continued to pay attention to the movies. ^Parameter^Value^ |Number of runs|3| |Volumes (e.g., TRs) per run|150| |''bio'' blocks per run|6| |''nonbio'' blocks per run|6| |''rest'' blocks per run|13| |Block duration|12 seconds| |Block order|''rest1'' => ''bio1'' => ''Rest2'' => ''nonbio1'' => ''rest3'' ... ''bio6'' => ''rest12'' => ''nonbio6'' => ''rest13''| For this project we will not explicitly model the presentation of the static movies. Because they occurred in both the bio- and non-bio- conditions, any activation they evoked will 'cancel out' in our contrast. ==== Data Acquisition Parameters ==== ^Parameter^Value^ |Field Strength|3.0 T| |Head Coil|12-channel| |Sequence|Echo Planar Imaging| |Martix|64 x 64| |Slices|37| |Field of View|22.4 cm| |Voxel size (x,y,z)|3.5 mm x 3.5 mm x 3.5 mm| |Orientation|Axial| |TR|2000 ms| |TE|25 ms| |Flip angle|90°| |Slice order|1,3,5...,37,2,4,6,...36| ==== Your Assigned Subject for First- and Second-level Analyses==== You are assigned a single subject (see below), who has three runs of data. Run 1st level analyses on all three runs of data. ^Name^Subject IDs^ |Lizzie|2545| |Jay|2553| |Srila|2554| |Scout|2585| |Julia|2731| |Laura|2585| |Caitlyn|2767| |Sarah|5770| |Katrina|2814| |Sam|2545| |Frank|2553| Data (anatomical and functional) for your assigned subject is located in ''/Users/hnl/Desktop/class/input/fmri/loc/data/nifti'' * ''SUBJ_highres.nii.gz'' * ''SUBJ_coplanar.nii.gz'' * ''SUBJ_motion_run1.nii.gz'' * ''SUBJ_motion_run2.nii.gz'' * ''SUBJ_motion_run3.nii.gz'' ===== Analysis ===== You will carry out [[http://www.andrewengell.com/wiki/doku.php?id=psyc410_s2x:fmri_part2#part_1preprocessing_preparatory_for_first_level_analysis|first-]], [[http://www.andrewengell.com/wiki/doku.php?id=psyc410_s2x:fmri_part3#part_1second_level_analysis|second-]], and [[http://www.andrewengell.com/wiki/doku.php?id=psyc410_s2x:fmri_part3#part_2third_level_analysis|third-]]level analyses on data from this experiment. * The first- and second-level analyses will be performed on a single subject (see your assigned subject above). * The third-level analysis will be performed across seventeen subjects. ==== First-Level ==== Data should be: * temporally highpass-filtered with a 100 second cutoff. * motion corrected. * corrected for slice scan-time. * spatially smoothed with a 5 mm kernel. * registered to standard space. * The hemodynamic response should be modeled with a Gamma-variate convolution model * Phase=0 * Stddev=3 * Mean lag=6 * Include contrasts for: * the main effect of biomotion * the main effect of non-biomotion * biomoition > non-biomotion * non-biomotion > biomotion * Do //**not**// include temporal derivatives in your model. First-level Output Directories Anatomical data: * ''~/Desktop/yourLastName_fmri_proj/SUBJ/SUBJ_coplanar_brain'' * ''~/Desktop/yourLastName_fmri_proj/SUBJ/SUBJ_highres_brain'' Analyses: * ''~/Desktop/yourLastName_fmri_proj/SUBJ/run01'' * ''~/Desktop/yourLastName_fmri_proj/SUBJ/run02'' * ''~/Desktop/yourLastName_fmri_proj/SUBJ/run03'' NOTE: You need specify the correct output directory for the runs, but ''FSL'' will create them. ==== Second-level ==== Perform a fixed-effects second-level analysis for your assigned participant. Results should be cluster corrected with a voxel z threshold of 2.3 and a cluster threshold of //p// < .05. Second-level Output Directory * '''~/Desktop/yourLastName_fmri_proj/SUBJ/allruns'' NOTE: You need specify the correct output directory for ''allruns'', but ''FSL'' will create it. ==== Third-level ==== In addition to your assigned subject, there are sixteen additional subjects. You will run a third-level analysis across all seventeen subjects. However, as with your previous lab, you do not have to do the first or second level analysis first as it has been done for you. Run a mixed-effects (FLAME1) third-level analysis. Results should be cluster corrected with a voxel z threshold of 3.1 and a cluster threshold of p < .05. Data are located in ''~/Desktop/class/input/fsl/loc/analysis/fsl'' in the following subdirectories: ''2545 2552 2553 2554 2585 2731 2766 2767 2814 3850 3851 3855 3866 5743 5744 5769 5770'' You should use the second-level analysis that //you created// for your participant. For all other participants you can use the second-level analysis that I created for you. Third-level Output Directories * ''~/Desktop/yourLastName_fmri_proj/third_level/cope1'' * ''~/Desktop/yourLastName_fmri_proj/third_level/cope2'' * ''~/Desktop/yourLastName_fmri_proj/third_level/cope3'' * ''~/Desktop/yourLastName_fmri_proj/third_level/cope4'' ===== fMRI Report ===== You will write-up your results a brief report in the style of a journal article, which includes the following sections. **//Introduction//** * Brief description (400-700 words) of the phenomenon under investigation (i.e., biological motion) and the goal of your investigation. * This will require that you do a little bit of research on biological motion. But given the nature of the project and the length of this section, your review of the literature can be fairly superficial (maybe find one or two papers that summarize or review the basic finding). * The goal of your study was to replicate the results of prior studies. Here's a brief (though somewhat old) review to get you started. {{:psyc410:documents:puce_2003.pdf|Puce & Perrett, 2003}} **//Methods//** * a description of the task paradigm * a description of your particpants * a description of the fMRI parameters (e.g., TR, slice order, flip angle, etc.). * a description of your preprocessing steps, including a sentence or two explaining why each step was performed. * a description of your statistical analysis, including the multiple comparisons correction you used Write the methods and results section as if you performed the first- and second-level analyses on all seventeen participants. For an example of how to write the fMRI methods and what to include, see {{:psyc410:documents:shultz_2014_social_cognitive_and_affective_neurosc.pdf|Shultz, van den Honert, Engell, & McCarthy, 2014}} * Use that paper as a guide. Do not simply copy from that paper or include information that is not relevant to this project. **//Results//** * A summary of your main findings with figures * Report which areas of the brain are selectively active (e.g., "//selective activation to faces was found along the bilateral fusiform gyrus and the right posterior temporal sulcus//") * Make your figures attractive and informative! * How you choose to display your data (e.g., on a 2D or 3D brain) should be based on what best highlights your results. * Show all relevant views * Label each figure (e.g., "Figure 1") and be sure to refer to each figure somewhere in the body of your text. In other words, don't just include figures and figure labels that are not referred to in the text. See the various example papers throughout this wiki (and those you find for your brief lit review) to understand how results are reported. While you //do// report the parameters used in various processing stages, you do //not// discuss the results of "first- or second-level" analysis. Indeed, those are terms specific to the ''FSL'' software and would be meaningless to a reader who happens to use ''AFNI''.. Don't forget to label which hemisphere is "R" and to include MNI coordinates. **//Discussion//** * Briefly discuss (400-700 words) your results with regard to the task * Compare your results to previous work, including [[http://www.andrewengell.com/wp/wp-content/PDFs/Engell%202013%20Neuroimage.pdf|Engell and McCarthy, 2013]]. **//References//** * Any citations and references should be in **__APA style__**. * See [[https://owl.english.purdue.edu/owl/resource/560/01/|here]] if you are unfamiliar with APA style. I would like the paper formatted like a professional manuscript. I have therefore made available a {{:psyc410:psyc410_fmriproject_formatting_template.doc|template manuscript}} from which you can work (this was ultimately published in the //Journal of Cognitive Neuroscience// in 2018). Of course, your paper does not need to be this long and so I've included reminders of your project's min/max word count at the start of each section.