1 / 63

Understanding fMRI Dynamics: Neuronal Influences & Hemodynamics

Explore the influences on fMRI signals, neuronal currents, and hemodynamic modeling in this study. Discover patterns, fluctuations, and experimental designs, and compare MEG and fMRI results.

lewismartinez
Download Presentation

Understanding fMRI Dynamics: Neuronal Influences & Hemodynamics

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 closer look at fMRI dynamics, fluctuations, and patterns Peter A. Bandettini, Ph.D. Section on Functional Imaging Methods Laboratory of Brain and Cognition, NIMH & Functional MRI Facility, NIMH

  2. 1991

  3. Measured Signal Neuronal Activation ? ? ? ? Hemodynamics Noise

  4. Dynamics • Fluctuations • Experimental Design • Pattern Information • Neuronal Current MRI

  5. Dynamics • Motivation: • To understand the neuronal and non-neuronal influences on the fMRI signal. • Studies: • Modulate timing: “on” duration, “off” duration, and duty cycle of visual cortex activation. • Neuronal and Hemodynamic Modeling. • MEG and fMRI Comparison.

  6. The BOLD Signal Blood Oxygenation Level Dependent (BOLD) signal changes task task

  7. Brief “off” periods produce smaller decreases than expected. measured 2 1.5 0 linear Signal (%) 1 Linearity -2 measured 0.5 linear -4 0 0 5 10 15 20 time (s) Stimulus OFF duration (s) R.M. Birn, P. A. Bandettini, NeuroImage, 27, 70-82 (2005) Brief “on” periods produce larger increases than expected. measured linear 3 Linearity 2 Signal linear 1 0 1 2 3 4 5 time (s) Stimulus Duration (s) R. M. Birn, Z. Saad, P. A. Bandettini, NeuroImage, 14: 817-826, (2001)

  8. Brief ON 25% 50% 75% Brief OFF 8% ON 25% ON 50% ON 2 75% ON Signal 1.5 Linearity 1 0 5 15 10 time (s) 0.5 0 0.5 1 Duty Cycle Varying the Duty Cycle Deconvolved Response R.M. Birn, P. A. Bandettini, NeuroImage, 27, 70-82 (2005)

  9. O2 Brief ON FIn FOut 25% 50% DV 75% Brief OFF Data 2 1.5 Linearity 1 0.5 0 0.5 1 Duty Cycle Simulation of Hemodynamic Mechanisms(Balloon model) E(f) = oxygen extraction fraction V = blood volume E(f)=NL , DV E(f)=NL, DV=0 E(f)=Lin, DV E(f)=Lin, DV=0 a d b c 0.4 0.8 0.8 0.6 0.6 0.6 0.4 0.2 0.4 0.4 BOLD Signal 0.2 0.2 0.2 0 0 0 0 0 5 10 15 0 5 10 15 0 5 10 15 0 5 10 15 time time time time e f g h 1.5 1.5 1.5 1.5 Linearity 1 1 1 1 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 0.2 0 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Duty Cycle Duty Cycle Duty Cycle Duty Cycle

  10. Brief ON OFF response Refraction 25% Adaptation 50% 75% Brief OFF Data 2 1.5 Linearity 1 0.5 0 0.5 1 Duty Cycle 0 5 10 15 0 5 10 15 0 5 10 15 0 5 10 15 2 2 2 2 1 1 1 1 Linearity 0 0 0 0 0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1 Duty Cycle Duty Cycle Duty Cycle Duty Cycle Simulation of Neuronal Mechanisms Adaptation + Refraction + OFF response Adaptation + Refraction Linear Adaptation 2 1 1 1 1 BOLD Signal

  11. MEG & fMRI Linearity Comparison A. Tuan, R. M. Birn, P. A. Bandettini, G. M. Boynton, (submitted)

  12. MEG Results A. Tuan, R. M. Birn, P. A. Bandettini, G. M. Boynton, (submitted)

  13. Measured and Predicted BOLD responses Non-linearity computed from MEG prediction 3 0s ramp 0.5s ramp 1s ramp Non-linearity 2 1 0 5 10 15 Stimulus Duration Non-linearity computed from BOLD 3 Non-linearity 2 1 0 5 10 15 Stimulus Duration (s) BOLDMEG A. Tuan, R. M. Birn, P. A. Bandettini, G. M. Boynton, (submitted)

  14. 2. Fluctuations • Motivation: • Applications of connectivity mapping (autism, schizophrenia, Alzheimer’s, ADHD) have exploded – need for better interpretation. • Distinguish neuronal activity-related fluctuations from non-neuronal physiological fluctuations. • -reduce false positives in resting state connectivity maps • -increase functional contrast to noise for activation maps • fMRI activation magnitude calibration using fluctuations rather than hypercapnic or breath-hold stress. • Studies: • Time course of respiration volume per unit time (RVT) • The Respiration Response Function (RRF) • FMRI Calibration using RRF

  15. Resting State Correlations Activation: correlation with reference function Rest: seed voxel in motor cortex B. Biswal et al., MRM, 34:537 (1995)

  16. BOLD correlated with SCR during “Rest” J. C. Patterson II, L. G. Ungerleider, and P. A Bandettini, NeuroImage 17: 1787-1806, (2002).

  17. Methodology Resting state networks identified with ICA M. DeLuca, C.F. Beckmann, N. De Stefano, P.M. Matthews, S.M. Smith, fMRI resting state networks define distinct modes of long-distance interactions in the human brain. NeuroImage, 29, 1359-1367

  18. Sources of time series fluctuations: • Blood, brain and CSF pulsation • Vasomotion • Breathing cycle (B0 shifts with lung expansion) • Bulk motion • Scanner instabilities • Changes in blood CO2 (changes in breathing) • Spontaneous neuronal activity

  19. 5 Z 0 5 % Breath-holding Group Maps (N = 7) MR Signal Respiration Cue time (s) 0 50 100 150 200 250 300 350 Breath-hold response (average Z-score) Anatomy R.M. Birn, J. A. Diamond, M. A. Smith, P. A. Bandettini, NeuroImage, 31, 1536-1548

  20. 50 100 150 200 250 300 350 1 0.5 CC 2 0 CO2 0 50 100 150 200 250 300 350 -0.5 RVT -20 -10 0 10 time (s) 0 Shift (s) 0 50 100 150 200 250 T max min 300 310 315 305 time (s) max - min RVT = T Estimating respiration volume changes Respiration time (s) Respiration Volume / Time (RVT) RVT precedes end tidal CO2 by 5 sec.

  21. 0 Z -4 5 Z 0 Respiration induced signal changes Rest 4 CC = 0.76 Rest (%) 2 Breath-holding 0 0 9 4.5 Breath hold (%) (N=7)

  22. 0 BOLD Z Respiration Cue 0 100 200 300 Time (sec) -4 5 Z BOLD Respiration RVT 0 0 100 200 300 Time (sec) Respiration induced signal changes Rest Breath-holding (N=7) R.M. Birn, J. A. Diamond, M. A. Smith, P. A. Bandettini, NeuroImage, 31, 1536-1548 (2006)

  23. Resting state correlation with signal from posterior cingulate Resting state correlation with RVT signal 10 6 |Z| Z 0 -10 RVT Correlation Maps &Functional Connectivity Maps Group (n=10) R.M. Birn, J. A. Diamond, M. A. Smith, P. A. Bandettini, NeuroImage, 31, 1536-1548 (2006)

  24. Effect of Respiration Rate Consistency on Resting Correlation Maps 10 Z -10 Spontaneously Varying Respiration Rate Constant Respiration Rate 10 10 Z Z -10 -10 Lexical Decision Making Task Blue: deactivated network Group (n=10)

  25. 0 50 100 150 200 250 300 350 time (s) 0 50 100 150 200 250 300 350 time (s) Respiration Changes vs. BOLD How are the BOLD changes related to respiration variations? RVT ? fMRI Signal

  26. fMRI Signal 1 D Signal (%) 0 2.1 3.54 t e t e RRF(t) = 0.6 0.0023 0 10 20 30 40 -1 time (s) t t 0 20 40 60 Deep Breath time (s) 4.25 1.6 fMRI response to a single Deep Breath Respiration … 40s deconv. Respiration Response Function (RRF) R.M. Birn, M. A. Smith, T. B. Jones, P. A. Bandettini, NeuroImage, (in press)

  27. 0 -3 0 100 200 300 Respiration response function predicts BOLD signal associated with breathing changes better than activation response function. 4 Breath-holding 2 Signal (%) 0 -2 20s 40-60s 0 100 200 300 time (s) Rate Changes Signal (%) 20s 40s time (s) Depth Changes 3 0 Signal (%) 20s 40s -4 0 100 200 300 time (s) R.M. Birn, M. A. Smith, T. B. Jones, P. A. Bandettini, NeuroImage, (in press)

  28. R.M. Birn, M. A. Smith, T. B. Jones, P. A. Bandettini, NeuroImage, (in press)

  29. %DS (BOLD) BOLDcalib = %DS (Resp) BOLD magnitude calibration Before Calibration After Calibration Respiration-induced DS Breath Hold Rest Depth Change Rate Change R.M. Birn, M. A. Smith, T. B. Jones, P. A. Bandettini, NeuroImage, (in press)

  30. 3. Experimental Design • Motivation: • Guides for individual subject scanning at the limits of detectability, resolution, available time, and subject performance. • Studies: • Overt response timing • Suggested resolution

  31. t fMRI during tasks that involve brief motion Blocked Design motion BOLD response task Event-Related Design BOLD response motion task R. M. Birn, P. A. Bandettini, R. W. Cox, R. Shaker, Human Brain Mapping 7: 106-114 (1999).

  32. Blocked min SD=9s SD=2s 20 min SD=3s min SD=1s min SD=7s min SD=5s 10 Detection of Motion (t-stat) SD=8s 0 Event-Related varying ISI Event-Related constant ISI (SD=1s, ISI=15s) 10 SD=4s SD=6s 20 0 10 20 30 40 50 60 70 80 90 100 Detection of BOLD (t-stat) Overt Responses - Simulations SD = stimulus duration More Motion Artifacts Better BOLD Detection R.M. Birn, R. W. Cox, P. A. Bandettini, NeuroImage, 23, 1046-1058 (2004)

  33. t-stat. 15 0 -15 Blocked design (30s/30s) Blocked design (10s/10s) Event-related Varying ISI (1s min. SD) Event-related Varying ISI (5s min. SD) Event-related constant ISI (1s. SD, 15sISI) Overt Responses

  34. 16 channel parallel receiver coil 8 channel parallel receiver coil GE birdcage GE 8 channel coil Nova 8 channel coil J. Bodurka, et al, Magnetic Resonance in Medicine 51 (2004) 165-171.

  35. Finding the “suggested voxel volume” Temporal Signal to Noise Ratio (TSNR) vs. Signal to Noise Ratio (SNR) 3T, birdcage: 2.5 mm3 3T, 16 channel: 1.8 mm3 7T, 16 channel: 1.4 mm3 J. Bodurka, F. Ye, N Petridou, K. Murphy, P. A. Bandettini, NeuroImage, 34, 542-549 (2007)

  36. Segmentation using EPI Transient J. Bodurka, F. Ye, N Petridou, P. A. Bandettini. NeuroImage, 34, 542-549 (2007)

  37. Sensitivity, Scan Time, and Temporal Signal to Noise K. Murphy, J. Bodurka, P. A. Bandettini, NeuroImage, 34, 565-574 (2007)

  38. 4. Pattern-Information Analysis • Motivation: • Classical fMRI analysis: • Is a region activated during a task? • Pattern-information analysis: • Does a region carry a particular kind of information? • Study: • Pattern-Information Mapping • Dis-similarity matrix

  39. Pattern Information Mapping “searchlight” ROI N. Kriegeskorte, R. Goebel, P. Bandettini, Proc. Nat'l. Acad. Sci. USA, 103, 3863-3868 (2006) From fixed ROI

  40. 96 0 0 0 0 0 0 dissimilarity matrix 96 0 0 0 0 0 0 96 Dissimilarity Matrix Creation compute dissimilarity (1-correlation across space) response patterns ... ROI in Brain stimuli ... N. Kriegeskorte, et al (in review)

  41. Visual Stimuli

  42. Human Early Visual Cortex(1057 visually most responsive voxels) Human IT(1000 visually most responsive voxels)

  43. Human fMRI in four subjects(repeated sessions,>12 runs per subject) "quick" event-related design(stimulus duration: 300ms,stimulus onset asynchrony: 4s) fixation task(with discrimination of fixation-point color changes) occipitotemporal measurement slab(5-cm thick) small voxels (1.951.952mm3) 3T magnet, 16-channel coil (SENSE, acc. fac. 2) Monkey (Kiani et al. 2007) single-cell recordingsin two monkeys rapid serial presentation(stimulus duration: 105ms) fixation task electrodes in anterior IT(left in monkey 1, right in monkey 2) 674 cells total windowed spike count(140-ms window starting 71ms after stimulus onset) Monkey-Human Comparison Procedure

  44. average of 4 subjects fixation-color task 316 voxels average of 2 monkeys fixation task >600 cells N. Kriegeskorte, et al (in review)

  45. 5. Neuronal Current MRI • Motivation: • Direct fMRI of neuronal activity. • Studies: • Model • Phantom Studies • Cell Cultures at 7T and 3T

  46. Magnetic Field Intracellular Current Surface Fields 100 fT at on the scalp J.P. Wikswo Jr et al. J Clin Neuronphy 8(2): 170-188, 1991

  47. Magnetic field associated with a bundle of dendrites Because BMEG=100fTis measured by MEG on the scalp at least50,000 neurons(0.002 fT (per dendrite) x 50,000 = 100 fT), must coherently act to generate such field. These bundles of neurons produce, within a typical voxel, 1 mm x 1 mm x 1 mm, a field of order: BMRI0.2nT

  48. Surface Field Distribution Across Spatial Scales Adapted from: J.P. Wikswo Jr et al. J Clin Neurophy 8(2): 170-188, 1991

  49. Current Phantom Experiment wire Z wire X

  50. B0 calculated Bc ||B0 Measurement 70 A current   200 Single shot GE EPI Correlation image J. Bodurka, P. A. Bandettini. Magn. Reson. Med. 47: 1052-1058, (2002).

More Related