A fMRI approach to probe CNS interaction Wei Chen, MD, MS. Modality : Animal MRI Mentor : Professor Seong-Gi Kim Kim’s Lab Faculty : Seong-Gi Kim, Tae.

A fMRI approach to probe CNS interaction

Wei Chen, MD, MS.

Modality : Animal MRIMentor : Professor Seong-Gi Kim

Kim’s Lab Faculty : Seong-Gi Kim , Tae Kim , Tao Jin, Hiro Fukuda , Alberto Vazquez.Lab Manager: Ping Wang

Functional MR imaging has been widely used to map brain functions. Due to slow hemodynamic responses, neural interaction between brain areas has not been easily investigated by fMRI.

This study was designed to examine neural interaction between two hemispheres by using bilateral somatosensory stimuli based on Ogawa et al. works (2000).

Background

In 2000, Ogawa et al. monitored rats’ brain BOLD response to left and right forepaw stimulation with the different interstimulus interval (ISI).

Seiji Ogawa, et al ,PNAS, 2000.

0 msec

ROI

Left Right

Right Hemisphere Simultaneously stimulate both sides’ forepaw

Seiji Ogawa, et al ,PNAS, 2000.

~40 msec

ROI

Left Right

Right -> Left forepaw stimulation

40 ms

75 ms

Right Hemisphere

Interaction between left and right hemisphere occurs when time interval between two stimuli is ~40 ms.

Based on the Ogawa et al. studies, this animal study was conducted to Reproduce the bilateral forepaw stimuli approach at 9.4 T.

Test the changed protocol Change anesthetic from alpha-chloralose (for terminal

studies) to isoflurane (for survival studies)

Use contrast agent, monocrystalline iron oxide nanoparticle (MION) to enhance the MRI sensitivity.

Task

Pulse width = 1.0ms, current = 1.2-1.5mA

R RLISI

LISI

1 sec

1 Hz

Left forepaw stimulation was given by following the right forepaw stimulation with the ISI = 0, 40 and 83ms.

Design (fMRI approach )

Right Left

Pulse generator Constant current isolator

fMRI measurement• 9.4-T MR system (Varian, Palo Alto, CA). • fMRI experiments: • Gradient EPI , TE =20 ms (11ms with MION), TR=500 ms• Data matrix = 64 x 32 (readout direction x phase-encoding

direction) • FOV= 32 x 16 mm (right--left hemisphere x dorsal--ventral

directions).• The power of the radio frequency pulse was adjusted to

maximize the signals at the targeted somatosensory cortex. • Number of slices: 3• Each run: Total 60-80 images with the following diagram

0

10-20 s10 s10 s

30-40 s

Postprocessing of fMRI data

Software: MATLAB & STIMULATE

Common approaches: parametric statistical methods such as the t test or cross-correlation (CC) . Use CC > 0.35 as a threshold.

1. Generate a BOLD map from a fMRI data set.

Image # (x 0.5 s)

1%

10%

BOLD fMRI Maps with ISI = 0 ms

Scan #1

Scan #2

Scan #3

2. Check ROI time courses of Scan #1

Left S1FL Right S1FL L R

Image # (x 0.5 s) Image # (x 0.5 s)

Inte

nsity

3. Check three scans’ Left S1FL baseline fluctuations

The reproducibility of fMRI signals is hampered because of trial-by-trial variations in baseline physiology (and in evoked responses).

Nor

mal

ized

In

tens

ity

Image # (x 0.5 s)

4. BOLD fMRI with ISI = 40 ms

1%

10%Scan #1

Scan #2

Scan #3

Left S1FL Right S1FL

The three scans’ average time course

Inte

nsity

Image # (x 0.5 s) Image # (x 0.5 s)

L R

5. Cerebral Blood Volume (CBV) weighted fMRI

In typical fMRI studies, neural activity induces BOLD signal increases. After the injection of contrast agents, negative signal changes were observed, mainly in tissue area. These changes are directly related to blood volume changes.

6. Representative average maps of seven scans

-10%

-1%

Left S1FL Right S1FL

ISI = 0ms

ISI = 40ms

ISI = 80ms

L R

Inte

nsi

ty

Image # (x 0.5 s)

Conclusions

1. The reproducible BOLD activation in S1 areas in this animal was not seen clearly. The reason can be that isoflurane reduces BOLD signal changes (T. Kim et al. 2010) at 1 Hz stimulation, which may not be detectable due to limited signal averaging.

T. Kim et al. / NeuroImage 52 (2010) 224–233

2. Subject physiology condition is very important for reproducible fMRI studies.

3. It is very important to understand the effect of anesthetic on hemodynamic responses for animal fMRI studies.

4. Contrast agents are very helpful for enhancing the MRI sensitivity.

Conclusions

What I have learned Rigorous examinations of fMRI data are essential.

It is very important to have a proper choice of the statistical methods.

Maintaining good physiological condition is very important for animal fMRI studies.

An active voxel may not necessarily mean that neuronal activity is present in that region, but may be due to artifacts including hemodynamic signals that do not exactly colocalize with the neuronal activity sites.

It is necessary to make more scans or more number of subjects to get the repeatable fMRI data.

Acknowledgements:

The MNTP Summer Workshop all the training faculty and TA(Drs. William Eddy, Mark Wheeler, Charles Laymon, Kwan-Jin Jung, et al )The MNTP Peers

Tomika Cohen, Rebecca Clark, 9.4 T Manager : Kristy Hendrich System Administrator and postdoc: Shafiq Abedin and Yuguang Meng