• Monika Pawłowska added an answer:
    What is the best technique for clearing the brain while preserving endogenously expressed GFP?

    I have looked at CLARITY, SeeDB, 3DISCO, in the literature and they almost always require immunolabeling. Which technique preserves GFP (or YFP) the best, without additional immuno?

    Monika Pawłowska · Nencki Institute of Experimental Biology

    Thank you for quick answers! Perhaps we should try RIMS too. Another person in my lab does immunostaining of slices after CUBIC, but we aim at thicker samples that are probably too thick for immuno.
    Mario: we've used glycerol already, but not the way you describe it, thanks for the link!

  • Jean-Pierre Thibaut added an answer:
    Does anyone have data on both features and exemplars of a concept combination and its constituents?

    Does anyone happen to have (or know of) data on both features and exemplars for a particular concept combination and also for its two constituents? (e.g. feature and exemplar data on PET, feature and exemplar data on FISH, and feature and exemplar data on PET FISH?) There is lots of partial data but we are looking for a complete data set of this kind. 

    Many thanks in advance. 


    Jean-Pierre Thibaut · University of Burgundy

    James Hampton (City university of London) has done a lot on the conceptual combination issue.
    Or Geert Storms in Leuven Belgium. I think that Geert has some norms on this. It sounds familiar.

  • Christos Papantoniou added an answer:
    Is there a way to in vivo visualize the formation of a memory?

    I was wondering if there was a way to "see" the activation of all the neurons needed to form a new memory in vivo. For example, to begin from the hippocampus and end to the cortical areas in which the memory would be stored.

    Christos Papantoniou · University of Groningen

    Thanks a lot Kees!

  • Matthias Preusser added an answer:
    Neuroimaging immunotherapy responses: what are your ideas on utilizing imaging techniques to track GBM response to immunotherapy?

    I am interested in developing methods that would be clinically relevant for imaging responses to immunotherapy for GBM. Any thoughts? 

    Matthias Preusser · Medical University of Vienna

     We have a paragraph on this topic in a recent review: 

  • Meghan O'Sullivan added an answer:
    Can anyone please explain why I might be finding FWE significant and not FDR in a ROI analysis?

    I am currently analyzing SPM results from a sample of 175. However, when investigating between group associations, I noticed that one of my results were significant for FWE-corrections, and not for FDR-corrections during ROI analysis. As I understand it, FDR corrections are more liberal, while FWE corrections are generally more conservative, so I cannot make sense of this finding- any ideas? 

    Meghan O'Sullivan · King's College London

    HI there,

    I am referring to peak-level inferences. And for the ROI analysis, an explicit mask was applied at batch-level... so yes, in a way this is a correction, but in order to report significance, I understand that I need to report the p-value using either FWE or FDR correction (selected from the output of whole brain analysis). 

    I hope this makes sense. Many thanks

  • Rakesh Pandey added an answer:
    Can anyone provide details of SynAmps 2 EEG amplifier with Curry Scan 7 Neuroimaging Suite?

    Esteemed RG members

    For a collaborative research project I wish to use SynAmps 2 64 channel EEG amplifier as the same device is proposed to be used at UK site of the project. For this amplifier the Compumedics suggest Curry Scan 7 neuroimaging suite and for ERP Stim2. My queries are:

    1. Is the said EEG system better of comparable to Biosemi Active-2 system?

    2. If I prefer to use SynAmps (to have parity in the equipment with other collborators) then can I use some open source /free software for data acquisition and analysis or I will have to use their proposed software?

    3. There are different licenses for Curry 7. Can  EEG and ERP studies be performed by the "acquisition and processing" license?

    4. What are the advantages of going for the license of "Basic, Advanced source and image analysis"?

    5.. For ERP studies they suggest Stim2 (stimulus presentation software). Can I use SuperLab or DirectRT or e-prime for the same?

    I am not well versed with EEG systems and EEG studies and thus a more detailed suggestion with helpful links will be of much help to me.

    With best regards

    Rakesh Pandey · Banaras Hindu University

    Thanks Ross and Rodolfo for your clear and comprehensive answers to my queries and helpful suggestions. From your replies it is apparent that Synamps2 will be very useful and helpful for my studies.

  • Juergen Ralf Schlaier added an answer:
    Is it really necessary to run two complete acquisitions with opposing polarities of the phase-encode blips with FSL TOPUP?

    Or is there any way to apply TOPUP with one complete acquisition (64 read-out directions) and only one anti-parallel B0 Volume to estimate the magnetic field and use "applytopup"?

    I came as far as to create the two files (which look quite good in fslview):



    But when I apply:

    "applytopup --imain=blip_up,blip_down --inindex=1,2 --datatin=my_acq_param.txt --topup=my_topup_results --out=my_hifi_images"

     the program tells me (understandably) that the two sets are mismatched and I get no results. Since I only have the 4D data set of the "blip_up" acquisition and one anti-parallel B0 volume, is there any way around that problem? If there is, how would I have to change the command line to obtain unwarped images of "blip_up" ?

    Juergen Ralf Schlaier · University Hospital Regensburg

    Hi, it´s me again.

    I have talked to our IT specialist,who would be really grateful if we could have parts of the script for your pipeline. That would make his and my life a lot easier. If it is more convenient for you to send it to me directly, this is my email account:

    Thank you again very much

  • Andy Yeung added an answer:
    Why such a huge difference between different VBM8 methods and which one should be selected?

    Hi everybody! 

    I have been using VBM8 in SPM12 for a few months now for volumetric analysis of the human Hippocampus, between two different groups. I'm completely aware that depending of the kind of modulation you apply to your data (affine vs non-linear) the different results you get. But in terms of correctness I didn't expect such a huge difference. Using the prior one (affine) and literately following the VBM8 manual I achieve very strange results being quite contradictory in comparison to the last named procedure. The major set-up difference between those two is the affine vs the non-linear selection and using the head size as covariate only for the affine method.

    I have been thinking the possibility that this huge difference could be due to my contrast matrix  being [-1 1 0 0 0] vs [-1 1 0 0] for the affine and non-linear method respectively. But it looks to be right for the two sample test as I have two groups and three covariates (age, head size (only for the affine) and HAMD.

    I would appreciate if any of you could make this clear to me.

    Thank you in advance!

    Best regards; Cecilio. 

    Andy Yeung · The University of Hong Kong

    Dear Cecilio,

    As VBM8 Manual states: Affine+non-linear allows comparison of the absolute amount of tissue, while Non-linear only allows comparison of the absolute amount of tissue corrected for individual brain sizes. Do the two groups of subjects have significantly different average brain sizes? If so, the outcome measurement may differ 'hugely'.



  • Basar Bilgic added an answer:
    Is there any international network (just like DIAN) targeting asymptomatic progranulin and tau mutation carriers?

    Neuroimaging including resting state fMRI.

    Basar Bilgic · Istanbul University

    Thank you for the answer.

  • Zhiqiang Zhang added an answer:
    Can anyone suggest how to set a maximum pixel cluster size as a threshold in REST toolbox for RS fMRI seed based analysis?

    Resting state fMRI - detection DMN with seed in hippocampus:

    I would like to set a maximum pixel cluster size as threshold to compare seed based analysis results of data sets made on different scanners and with different head coils. This is not possible with the REST toolbox. Any other suggestions? Thanks in advance.

    Zhiqiang Zhang · Nanjing University

    Multiple correction for result representation?

    AlphaSim correction can be done on toolkit of REST: Utility-->REST AlphaSIM, which is not dependent on what's the analysis you done. seed-based FC, ALFF, task-state activation or else.... why emphaaaaasize RS fMRI seed based analysis?

  • Erik Bojorges added an answer:
    How can I minimize AC 50 Hz noise when recording EEG at the ICU?

    I understand that unplugging all sources from the AC supply helps, but there are several devices that must be plugged, such as:

    1. There are multiple devices pluged the the AC supply necessary for the EEG (& ERPs): two laptop computers and the EEG amplifier set.

    2. There are multiple devices pluged the the AC supply necessary for the patient: ventilator, intrevenal drug supply, several monitors for EKG, oxygenation, etc.

    Therefore, I wonder if there are ways to electrically insulate all cables involved in the EEG recording, as well as the EEG amplifier set.

    Erik Bojorges · Metropolitan Autonomous University

    congrats Efthymios!

  • Ulrich Mutze added an answer:
    What is the relationship between HSB brightness scale and luminance range of the screen?

    I have a following problem - I've measured with the use of manual photometer luminance of the screen expressed in cd/m2. Measurement was performed for RGB(0,0,0) and RGB(255,255,255) patterns. Thus, I've got luminance values equal respectively to 0.639 and 101.2 cd/m2. Now, taking into account resulting range created of those two values, I want to create 10 gray-scale colors equaly spread according to aforementioned luminance range, so I'll get following approximate luminance values of those colors:

    [ 10.0561 20.1122 30.1683 40.2244 50.2805 60.3366 70.3927 80.4488 90.5049 100.561 ]

    I was thinking about using HSB brigthness scale, but I have no clue how can I achieve values like shown above by manipulating aforementioned scale. For example, 50% brightness does not result in color luminance equal to ~55 cd/m2

    Thank you in advance for advice.

    Ulrich Mutze


    instead of your polynomial fit (which works not too bad in your case, I agree) the experts in the field use an exponential fit e.g. const B^gamma where it is tradition to name the exponent 'gamma'. This is remarkably accurate for many types of electrically driven radiation sources. I think it would be advantageous for you to have a closer look to the material which Curtis put together in his contribution (or to similar sources).

  • Marcie Zinn added an answer:
    Better than EEG?
    Does anyone know of any low-cost methods for non-invasively monitoring brain function, but with yielding more information, more robustly than EEG?
    Marcie Zinn · the NeuroCognitive Research Institute, United States.

    Learn to use eLORETA, Pascual-Marqui's source localization method. Go to his website and start reading. All you need is 30 seconds per subject of 19-channel EEG (or more channels).

  • Tae Young Lee added an answer:
    FSL 5.0 GUI working fine but Command line not working - any suggestions?
    I have installed FSL 5.0 in my ubuntu 12.10. It's working fine when processed from GUI but when I try to run it using command line, it shows "command not found".
    Tae Young Lee · Seoul National University
    Check both .bashrc and .bash_profile
  • Mark S Bolding added an answer:
    If you could buy any 3T human MRI right now, what would you get? Why would you choose that system?

    Note: this question was originally asked in 2012. I appreciate updates with recent information. 

    Ingenia, Achieva, Skyra, Verio, Discovery, which one is best for fMRI, for DTI, for MRS? How about a Trio with the new Tim 4G electronics? Or maybe you want a Fonar...

    Mark S Bolding · University of Alabama at Birmingham

    I received a lot of good feedback from this RG question. It led me down a long path which has ended in a very interesting way: by next year our institution will have recent 3T scanners from GE, Philips, and Siemens. We should be in a good position to do some direct comparisons.

    Basically, before you purchase a system, test your specific protocols on that specific system at a site that has used the scanner for a while and also get sample data from the factory. Bear in mind that there are things you can do at one field strength that you cannot easily do at another. That being said, here are some musings based on my search for the right scanner.

    It turned out to much harder than I expected to compare scanner performance or to answer a seemingly simple question like "Which scanner is better?" After talking to a lot of people and playing with a lot of scanners, here is my subjective opinion and some observations:

    GE, Philips, and Siemens all have great human MRIs. You can probably forget the other vendors unless you need something for animals or a special application (for many special applications, Bruker is wonderful). As long as you get a recent model scanner from one of those vendors, there are several other factors that are far more important than the choice of vendor. (1) Have excellent people to use the scanner, the best you can get. An MRI is like a musical instrument: you want a good one, but the person playing it is what really matters. (2) Setting up the facility to make it easy for investigators to do their work and facilitating development work is also much more important than which scanner you get. Ease of access and reducing administrative burden is surprisingly important for good science. (3) Building a good relationship with the vendor so that you have access to engineers and can talk to them directly is very important. Sometimes you need information or need to fix something and there is no other way but to talk to the person who designed that part of the machine or wrote the code. Ideally, you would want an applications person from the company on site as much as possible.

    In the context of research, and wanting a machine that is general purpose, a Siemens is a good choice. It is also a popular choice, which make multi-site studies easier. And, in my experience I have found them to be the easiest vendor to work with. (However, that may change when we start using the new GE PET/MR.) In general, I have found Philips to be difficult to work with. Though I will say, they were very helpful in fixing a problem with a fatsat module a few years ago. And, the prettiest MR images I have ever seen from a 3T scanner were made by my colleague Larry ver Hoef on our oldest Philips 3T. They were exquisite high res images of human hippocampus (in vivo).

    For 'DTI' (i.e. HARDI, DSI, etc.), get a Prisma.

    Apparently GE has a bad reputation among some (many?) research neuroimagers based on their older products and the way they treated research relationships. However, it seems things are changing for the better in both domains.

    Finally, here is how things turned out at our institution. I posted this question because we needed to replace an aging, head only, research dedicated MRI. We decided on a Siemens Prisma and it will be installed late this year. In a separate project made possible by a generous donor, we purchased a 3T GE PET/MR system which is based on a 750w platform. It is being installed now. It doesn't look like much on the outside, but on the inside it is like science fiction. The only problem is figuring out what to do with a PET/MR once it is running. Finally, we already had Philips Ingenias on the clinical side, which we use do use for research. They work well for what we do.

    I intend to update this thread after we have some experience with the new scanners. 

  • Ines Mürner-Lavanchy added an answer:
    What is the difference between small volume correction and ROI analysis in SPM?

    Dear all,

    I've been using ROI analysis for fMRI data analysis for a while, to search for only regions of interest supported by previous data or literature.

    But SPM has a SVC button with a similar logic behind. So what are the differences between the two approaches? Thanks a lot.


  • Bernd Schmeikal added an answer:
    What are the advantages and what are the problems of the hypothesis about “retinoid system”?
    Is the hypothesis about “retinoid system” (by Professor Arnold Trehub), as described in “Consciousness and Cognition” 16 (2007) 310–330 and in the other works of Professor Trehub, a plausible hypothesis? What are its advantages and what are its problems?

    Professor Trehub describes it in the following words:
    “Activation of the brainʼs putative retinoid system has been proposed as the
    neuronal substrate for our basic sense of being centered within a volumetric
    surround –- our minimal phenomenal consciousness (Trehub 2007). Here, the
    assumed properties of the self-locus within the retinoid model are shown to
    explain recent experimental findings relating to the out-of-body-experience. In
    addition, selective excursion of the heuristic self-locus is able to explain many
    important functions of consciousness, including the effective internal
    representation of a 3D space on the basis of 2D perspective depictions. Our
    sense of self-agency is shown to be a natural product of the role of the heuristic self-locus in the retinoid mechanism.” (Abstract, from: Where am I? Redux.)

    For the publications of Professor Trehub see:

    The question has been already discussed on the folowing thread:
    Bernd Schmeikal · University of Vienna

    Yes, Arnold, I think that Herb's work is most relevant, as he is familiar with the early works on cognitive developmental psychology and relations of genetic structuralism to quaternions.

  • Jiahao Ye added an answer:
    Is there any neuronal marker which labels neurons are inhibited rather than those are activated?

    I am aware of that you need right proportion of excitation and inhibition to get the brain works properly. Some neurons are activated when the animal performs certain behavior (i.e. fear learning) or cognition or emotional state, but there are also neurons which are inhibited in the meantime. We can have clues about those activated neurons by immediate early gene, like c-fos, Arc. How can you find those neurons who have their activities inhibited ? Does anyone has any clues or comments on this?

    Jiahao Ye · University of Strasbourg

    Hi, Damian& Francesco, yes I totally agree with quiescent neurons, most likely, never live in this planet. For a neuron, some of its activity can be attributed to background noise, and some caused by external stimuli which makes it fires more or fires less, right? If expressions of some genes subjected to certain level of elevation of activity, there is no reason that there is no genes are subjected to a certain level of inhibition of activity. If yes and the change is significant enough, then it can be neuronal marker for those neuron in 'state of inhibition'. As I noticed, in disease conditions like major depression mouse model, mPFC are inhibited.

    The dualism of activation and inhibition at the cellular level seems not a good idea when one tries to answer what are going on at cellular level during macroscopic behavior. One of the reasons is, as Francesco put it, the physiological diversity and the expression of so-called activity markers are results of a huge collection of factors. But to map the activation pattern at cellular level, IEG like c-fos helps a lot, though not ideally, in identifying neuronal circuit underlying specific behavior. These two concepts also play a lot in the optogenetics.

    But the discussions till now are quite interesting as we try to understand what really happen to these neurons. Thank you all!

  • Humera Tariq added an answer:
    Does anyone have any suggestions for brain MR image segmentation?
    I used T1 weighted brain MR images with 3mm slice thickness, 3% noise, and 40% intensity non-uniformity for segmentation into white matter, gray matter and CSF. Can I use Brain Web Discrete or fuzzy anatomical model as a gold standard? Why or why not? What is the best choice to compare different segmentation algorithms otherwise?
    Humera Tariq · University of Karachi

    Dear Coupe , thanks for  comments and suggestion. I'll definitely use your system for comparative analysis of tissue volume calculations.

  • Rahimi Ali added an answer:
    Do you know a good online course on how to use Psychopy software?
    I used Psyscope X during the last years but now I’m obliged to change software as our fMRI scanner cannot be interfaced with a mac but only with a PC. So I need to learn a new software that should be as easy as possible and allows me to trigger devices like TMS, EMG, etc. I think a good solution for me could be Psychopy as it is freeware and it's also a cross-platform software, so that it allows one to program using a Mac and then upload the software in the fMRI-PC. So the question is, is there any online course or podcast to learn step by step how to use it? This is so important to me that I will also consider using another software if it has a very good course. Thanks
    Rahimi Ali · Bangkok University

    Dear Giuseppe

    here it is :

    good luck with your research


  • Denver Ncube added an answer:
    What is the best way to do structural connectivity analysis? Immunofluorescence of radiological imaging?

    i am investigating neurodegeneration in different pathological states so i would wish to also do structural connectivity analysis.

    Denver Ncube · University of Zimbabwe

    George Nyandoro that would be great indeed, are there specific models that have developed in similar fashion which we could use?

  • David Clunie added an answer:
    Does anybody know where to find an online dataset of segmented brain CT Scans?

    I am studying automatic segmentation of pathologies in the brain, for which I would need a manually segmented and labelled set of CT brain scans. Does anybody know of such a segmented dataset?

    David Clunie · PixelMed Publishing, LLC

    The Harvard SPL Brain Atlas (MRI) might be useful, since it consists of image files and data sets that are usable by software, rather than being an "application" with a user interface. See "". I have played around with converting these from a Slicer-specific to standard DICOM encoding, as DICOM segmentation objects, so if you need me to explain how to do that, let me know ( Obviously for an MRI based atlas to be relevant to CT you would need to perform some kind of registration (but than you need to do that from a patient to an atlas anyway, regardless of modality).

  • Hayder Dibs added an answer:
    How can I apply deconvolution analysis, using BrainVoyager?

    The steps in BV are attached

    However, I don't really understand the output from BV...

    Lets say I did a fast event related design of 2 conditions: faces and objects.

    1. following this guide, after applying deconvolution GLM (5):
    A. Can i add the head motion to tthe new Deconvolution Design Matrix?

    B. should I run the contrast face / objects only to the the peak of the haemodynamic response?(i.e., if I chose 10 points for the FIR, for the contrast I should choose only 4-8)
    I don't really understand how BV calculate this GLM. Is it first find the shape of the hdf of the conditions (face and object) to each voxel, and then apply GLM?

    C. should i correct the statistic for a group ? (as there are much higher number of degrees of freedom).
    Furthermore, I artificially increase the degrees of freedom- from 2s to a virtual resolution of 1 s- how to correct the statistic?

    2. Running a ROI deconvolution analysis- i chose “ROI GLM” , and get 2 plot of the betas, 

    A. The second plot-  after AR(2)- i thought the correction for serial correlations affects only the error (not the beta).

    B. I thought the correct way to look at the event related averaging in % change (then calculate the peak or AUC). can i do it with bv?

    any advice will be much appreciated

  • Bob Jacobs added an answer:
    Can you suggest studies about what's going on into the brain while using media?

    I'm interested in studies concerning the relationship between neurosciences and media, such as what happens in our brain while we are using any type of media or technology.

    A good example is the work of Anderson (2007): "A neuroscience of children and media?" or the work of Bavelier on videogames. More references you may know?

    Thank you in advance.

    Bob Jacobs · Colorado College

    Good book I have used in class about this:

  • William Von Witt added an answer:
    Why does my DTI_FA.nii.gz file show blank mask in FSL?

    I've completed a DTI analysis with FSL before (tutorial data) with no issues. However, when I use my labs data, FSL won't produce a proper output. Literally a blank white mask is shown when I enter DTI_FA into FSL View. When I enter the original unprocessed nii.gz image I see a normal brain scan.

    I used this tutorial:

    I downloaded the DCM files from my server (ep2d_DWI) scan and converted them to FSL_4D files with dcm2niigui giving me bvec and bval outputs as well.

    I'm wondering if my issues are with the bvals, as it only has 3: '0,500,1000.' Where usually the tutorial files have 0, 1000 (repeated 20 or so times). In FSL View I see the following information when I only input the 4D file converted from DCM. 

    Voxels: 128x128x24

    Dimensions: 1.71875x1.71875x6mm

    Volumes: 3

    Data type: signed short (16bpp)  

    Any Ideas??

    William Von Witt · University of Queensland

    I like to use both MRTrix
    to look at new data, or rather to look at the info it contains with MRInfo, then to convert to whatever file type I need using MRConvert.  You can use it to create a brain mask too, but we found the brain mask tool in FSL created a more continuous mask with less chance of islands appearing. 
    You could try using MRInfo to have a look at the file info and make sure you are looking at the file you think you are.

  • Paul Summers added an answer:
    What is the best freely available software in order to study aneurysm growth based on Magnetic Resonance Time of Flight images ?

    I want to study aneurysm growth (in volume) in a set of subjects, based on Time of Flight images (MR) images at two time points.

    For each subject, I want to do a rigid registration between time of flight images as well as a segmentation of the cerebral vascular system. And then, I want to compare the two registered segmentation volumes in order to quantify the aneurysm growth.

    For doing this, some teams used in-house registration tools (not distributed as far as I know, AnToNIA for ex) and the registration is performed using a volume of interest around the aneurysm sac (and not using the whole brain).

    I currently know free registration tools not specific to aneurysms (FLIRT of FSL for exemple: Additional question: Do you think these kind of methods is appropriate to do "aneurysms registration" ? 

    Thanks in advance,

  • Oliver Grimm added an answer:
    How do I manually convert segmented images from BRAINS2 to NIFTI file format?

    manual segmentation done with the BRAINS2 toolbox gives two files describing the manual delineation in two planes. This gives two files with the ending "xroi" or "zroi". How do I convert these ROIs to NIFTI? Would be sad to loose these masks just because BRAINS2 is now outdated ...

    Oliver Grimm · Psychiatrische Universitätsklinik Zürich

    Vince Magnotta from Iowa answered by email my question in a very useful way. So maybe for some folks it's useful, too. You definitely need an installed version of BRAINS2. There are two ways. First, simply by using the GUI: 

    If you convert the ROIs to a mask and then you can convert the mask to an image. Both options should be available via the toolbar. You can then save the image out in analyze format which is the older version of the Nifti format

    An second by scripting. Every action in BRAINS e.g. clicking in the toolbar can be descibed in TCL. So automatizing would lead to something like:

    set r1 [b2 load ROI ${filename}]
    set m1 [b2 convert ROI to Mask $r1]
    set i1 [ b2 sum masks [ list $i1 ] ]
    b2 save image ${outfilename} strictAnalyze75 $i1

    Thank you Vince Magnotta!

  • David Skaer added an answer:
    Will an understanding of the neural substrates of mental disorders revolutionize diagnostics and treatment?

    Dear all,

    There is an expectation that understanding the neural substrates of mental disorders will revolutionize the field of psychiatry, in terms of improving diagnostics and treatment. However, the clinical value of neurobiological research into mental disorders has so far been minimal.

    What are your opinion on this matter?
    Do you agree that an understanding of the biology of psychiatric disorders (genetics, molecular, brain circuits etc) will lead to improved diagnostics and treatment, or are the expectations of a ”revolution” exaggerated?

    I’m interested in hearing your opinions on this matter.

    David Skaer · Saint Leo University

    If my "depressive thinking" can affect my neural activity (as some research indicates), then looking at the neural activity in such a case is not as helpful as straightening out my thinking.

    I've found so many "biological causes" of depression that are almost unrecognized as causes (e.g., influenza, Cushing's, etc), that one gets disheartened. In the 80s, I had many doctors question me why I thought hypothyroidism could be a cause. Fortunately, that is accepted pretty much nowadays. When I worked in a psychiatric hospital, the doctors were surprised that B-12 deficiency could mimic schizophrenia and senility...until they saw the miracles of B-12. I still find doctors not checking for some simple, biological causes of mental problems.


Topic followers (17,388) See all