1 / 28

Outline

A Cortical Area Selective for Visual Processing of the Human Body by P.E. Downing, Y. Jiang, M. Shuman, N.Kanwisher presented by Ilya Levner. Outline. Functional neuro-imaging Issues fMRI basics Data analysis with SPM Experiment Summary Conclusions. Integration vs Segregation.

kenna
Télécharger la présentation

Outline

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. A Cortical Area Selective for Visual Processing of the Human Bodyby P.E. Downing, Y. Jiang, M. Shuman, N.Kanwisherpresented by Ilya Levner

  2. Outline • Functional neuro-imaging • Issues • fMRI basics • Data analysis with SPM • Experiment Summary • Conclusions

  3. Integration vs Segregation • Are physiological changes elicited by sensorimotor or cognitive challenges best characterized by • Functional Segregation • the activation of functionally segregated areas • Functional Integration • Anatomically distributed, but functionally integrated system

  4. fMRI Basics • Based on BOLD Contrast • Blood Oxygenation Level Dependent contrast. • Deoxyhemoglobin is more paramagnetic (magnetizes easier) than oxyhemoglobin • So deoxyhemoglobin is the contrasting agent revealing cortical areas metabolizing oxygen at highter rates (presumably).

  5. fMRI (Cont) • Small contrast difference oxygenated and deoxygenated blood. • 1.5T fMRI produce image intensity changes of no more than 2-4%. • 2T fMRI produce image intensity changes of less than 15%. • 4T fMRI has produced contrast changes of up to 30%

  6. Statistical parametric map (SPM) Data Analysis Overview Design matrix Time-series data Kernel Realignment Smoothing General linear model Gaussian field theory Statistical inference Normalisation p <0.05 Template Parameter estimates

  7. Reasons for Motion Correction The Steps in Motion Correction • registration - i.e. determining the 6 parameters that describe the rigid body transformation between each image and a reference image. • transformation - i.e. re-sampling each image according to the determined transformation parameters. • Subjects will always move in the scanner. • movement may be related to the tasks performed. • The sensitivity of the analysis is determined by the amount of residual noise in the image series, so movement that is unrelated to the task will add to this noise and reduce the sensitivity.

  8. Spatial Normalisation Spatially normalised Original image • Enable reporting of activations as co-ordinates within a known standard space • e.g. the space described by Talairach & Tournoux • Inter-subject averaging • extrapolate findings to the population as a whole • increase activation signal above that obtained from single subject Spatial Normalisation Template image Deformation field

  9. Smoothing • Why Smooth? • Potentially increase signal to noise. • Inter-subject averaging. • Increase validity of SPM. • In SPM, smoothing is a convolution with a Gaussian kernel. • Kernel defined in terms of FWHM (full width at half maximum). Before convolution After convolution Gaussian smoothing kernel

  10. Statistical Parametric Maps (SPM) • Construction of statistical processes to test hypotheses about regionally specific effects. (functional segregation hypothesis) • Analysis of each voxel done using any standard (univariate) statistical test. • Voxel is a 3-D volumetric pixel.

  11. Voxel by voxel statistics… model specification parameter estimation hypothesis statistic statistic image or SPM f MRI time series voxel time series

  12. Model (green and red) Fitting (blue : global fit) Noise Correlation between regressors True signal

  13. …fitted raw fMRI time series adjusted for global & low Hz effects fitted box-car scaled for global changes fitted “high-pass filter” residuals

  14. Data Analysis Overview Statistical Parametric Map Design matrix fMRI time-series kernel Motion correction smoothing General Linear Model Spatial normalisation Parameter Estimates anatomical reference

  15. p = 0.05 p = 0.0001 p = 0.0000001 Simple threshold tests… 5mm 10mm 15mm

  16. Experimental Results

  17. Results (cont)

  18. Results Response of the extrastriate body area Ex-1) Human Body Parts 1.3%, face parts 1.0% vs object parts 0.5% Ex-2) hands 1.4% = body parts 1.4% Ex-3) Whole body 1.9% vs body parts 1.4% Ex-4) Ex-5) objects = object parts = 0.5% Ex-6) Stick figure 1.7%

  19. References P.E. Downing et al (1995): A Cortical Area for Visual Processing of the Human Body.SCIENCE, vol 293, pp 2470-2473. Friston et al (1997): SPM Short Course. Course Notes, 1997. Ashburner et al (1999): Nonlinear spatial normalisation using basis functions.Human Brain Mapping 7(4):254-266 Ashburner & Friston (2000): Voxel-based morphometry - the methods.NeuroImage 11:805-821

  20. fMRI Basics (Pros & Cons) • Pros • Spatio-temporal scale of 1-3mm at 1-2 seconds compared to PET with 6mm and 30 seconds. • Cons • Blood occupies only 2% of cortical mass. • Precision can’t get any finer. • Increase in blood flow extends well beyond the area of activation. • A cortical region can activate in 10ms while a change in blood flow can take up to 1s.

  21. Residual Errors from fMRI • Gaps between slices can cause aliasing artefacts • Re-sampling can introduce errors • especially tri-linear interpolation • Ghosts (and other artefacts) in the images • do not move according to the same rigid body rules as the subject • Slices are not acquired simultaneously • rapid movements not accounted for by rigid body model • fMRI images are distorted • rigid body model does not model these types of distortion • Spin excitation history effects • variations in residual magnetisation • Functions of the estimated motion parameters can be used as confounds in subsequent analyses.

  22. Spatial normalisation • Inter-subject averaging • extrapolate findings to the population as a whole • increase activation signal above that obtained from single subject • increase number of possible degrees of freedom allowed in statistical model • Enable reporting of activations as co-ordinates within a known standard space • e.g. the space described by Talairach & Tournoux • Warp the images such that functionally homologous regions from the different subjects are as close together as possible • Problems: • no exact match between structure and function • different brains are organised differently • computational problems (local minima, not enough information in the images, computationally expensive) • Compromise by correcting for gross differences followed by smoothing of normalised images

  23. T1 Transm T2 T1 305 T2 PD SS PD PET EPI Template Images “Canonical” images Spatial normalisation can be weighted so that non-brain voxels do not influence the result. Similar weighting masks can be used for normalising lesioned brains.

  24. Voxel-Based Morphometry Preparation of images for each subject Spatially normalised Partitioned grey matter Original image Smoothed A voxel by voxel statistical analysis is used to detect regional differences in the amount of grey matter between populations.

  25. General Linear Model… • fMRI time series: Y1 ,…,Ys ,…,YN • acquired at times t1,…,ts,…,tN • Model: Linear combination of basis functions • Ys = 1f 1(ts ) + …+lf l(ts ) + … +Lf L(ts ) + s • f l (.):basis functions • “reference waveforms” • dummy variables • l: parameters(fixed effects) • amplitudes of basis functions (regression slopes) • s: residual errors: s ~ N(0,2) • identically distributed • independent, or serially autocorrelation (Generalised Linear Model  GLM)

  26. ^ a (x2, Y2) Y 1   (x3, Y3)  (x1, Y1) ^ m x ^ ^ Y1x11ae1 Y2 = x21 + e2 Y3x31me3 Y = Xb + e ^ ^ ^ Example: a line through 3 points… simple linear regression parameter estimates m&a fitted values Y1 ,Y2 ,Y3 residuals e1 , e2 , e3 Yi = axi + m + eii = 1,2,3 Y1 = ax1 + m × 1 + e1 Y2 = ax2 + m × 1 + e2 Y3 = ax3 + m × 1 + e3 dummy variables

  27. fMRI box car example… parameters error vector design matrix data vector a m 3 4 5 6 7 8 9 = + Y = X  + 

  28. References Friston et al (1995): Spatial registration and normalisation of images.Human Brain Mapping 3(3):165-189 Ashburner & Friston (1997): Multimodal image coregistration and partitioning - a unified framework.NeuroImage 6(3):209-217 Collignon et al (1995): Automated multi-modality image registration based on information theory.IPMI’95 pp 263-274 Ashburner et al (1997): Incorporating prior knowledge into image registration.NeuroImage 6(4):344-352 Ashburner et al (1999): Nonlinear spatial normalisation using basis functions.Human Brain Mapping 7(4):254-266 Ashburner & Friston (2000): Voxel-based morphometry - the methods.NeuroImage 11:805-821

More Related