NetForum uses cookies to ensure that we give you the best experience on our website. If you continue to use the site, we'll assume that you are happy to receive these cookies on the NetForum website. Read about our cookies.
NetForum Community
Learn. Share. Optimize.
Log in | Sign up now | Submit content | Contact

Attention valued NetForum members:

Due to evolving technology needs and global privacy regulations, we have made the hard decision to suspend the NetForum User Community platform on Monday, December 9, 2019.

After this date, the current NetForum can no longer be reached. Please click here for more information about this decision, what happens next and how to stay in touch with us about the future of the community.

Go to similar content

10 tips for BOLD imaging

Application Tip
van den Heuvel, Martina Philips Healthcare
Hoogenraad, Frank, Ph.D. Philips Healthcare

BOLD imaging helps identifying active regions of the brain relying on local metabolic and hemodynamic changes which occur in activated cortical brain.


During brain activation (increase of metabolism), the oxygen consumption of local tissue increases by approximately 5% whereas the oxygen tension will decrease. The effects of local increases of microvascular oxygenation can be visualized.


This contrast mechanism is the BOLD effect where BOLD stands for "Blood Oxygen Level Dependent".

Tip 1: About paradigms

In general in BOLD imaging, instructions need to be given to the patient. These instructions are focused on delineating a certain area of brain function. As an example, the instruction could be to move the right thumb or to hold still. In this case, the instruction would be "move" or "hold still". This set of instructions is called a paradigm.


A Paradigm is the pattern of how brain activation is performed:

  • either a stimulus is presented
  • or a task is executed. 


The paradigm consists of blocks (of fixed duration) of brain activation alternating with blocks of rest (or with a control task). These blocks correspond to a specific 'number of dynamics'.

The paradigm is characterized by the length of activation and rest in dynamics. It may be more complicated consisting e.g. of different tasks, even of different intensities with varying length of the activation or rest blocks. 



ApTip RT BOLD imaging
ApTip RT BOLD imaging

Schematic display of a fingertapping paradigm. 

The paradigm consists of 10 dynamic scans in rest followed by 10 dynamics of left hand tapping and by 10 dynamics of right hand tapping. In general this paradigm is executed several times within a scan.

Tasks and stimulation

Brain activation can be detected with a very wide range of paradigms. Simple paradigms can be visual (different hemfields, flickering or static, moving or static), silent- or overt speech to identify Broca's and/or Wernicke's area, depicting motor activity of the different hand digits, listening to auditory stimuli etc.

More advanced paradigms focus on memory (by showing "new" versus "old" faces and places), attention (the "no-go" taskes) etc.


Motor tasks:

The patient is asked to perform a motoric task e.g. to tap his/her index finger.


Visual stimulation:

The patient is confronted with visual stimulation e.g. a flashing checkerboard.


Auditory stimulation:

The patient is confronted with auditory stimulation e.g. auditory presentation of words.


Tip 2: About the Functional Brain Imaging (FBI) box

The paradigms - and instructions - must be synchronized with the scanner. This means that the first instruction should be given at dynamic 1, the second instruction e.g. at dynamic 11 etc. This asks for precise work and a secure workflow.


This can be facilitated by the use of the FBI box which generates trigger pulses (TTL pulse of 4 V of around 5 microseconds) for an external device to produce the next instruction for the patient, e.g. give the next track of a mini-disk, or show the next slide on screen.


The trigger signals are derived from the corresponding imaging parameters "Start at dyn" and "Interval (dynamics)" which appear when "Synch. external device" has been set to "Yes".  


In such a way, the paradigm is performed automatically and without any sources of error.



Visual stimulation is presented to the patient via a projection screen in the magnet room. The image on the projection screen results via a beamer from a laptop which is connected to the FBI box and receives the trigger signals as input.


Each time the (next) trigger is shown the next visual input is given (e.g. by the next powerpoint slide).


Note that the trigger is given at the first RF pulse of the first slice of each dynamic/volume. So if the trigger is programmed (in the scan protocol) to provide a trigger at dynamic 5, it will be given after dynamic 4 has finished, and the first RF pulse has started for dynamic 5.


Tip 3: About the BOLD imaging sequence

In order to visualize the BOLD effect, a heavily T2*W scan has to be acquired.


In BOLD imaging, typically single-shot long-TE FFE measurements are performed with matrices of around 64 - 96. This will yield a TR or around 2 s to 4 s depending on the number of slices used.

Scan times are in the range of 4 min. to 5 min. with typically 25 slices, 80 dynamics of 3 s each, e.g. 10 dynamics in rest, 10 dynamics during activation, to be repeated 4 times.


Also the 3D PRESTO is often used for BOLD imaging (utilizing phase navigation). With this type of acquisition high temporal resolution can be produced by applying SENSE-factors in both phase-encoding directions.  

Tip 4: About patient preparation

  • Instruct the patient carefully about what to expect and about the paradigms he/she has to perform.


  • Tell the patient what he/she has to do, e.g. "Move the right thumb", and practice.  


  • Tell him/her also what is not desired, e.g. "Do not move the whole hand or even the arm" as this might give a stimulus correlated activation which may influence the funcional result.  


With visual stimulations, ensure that the patient can see the required information on screen by means of a mirror on the coil.


Tip 5: About planning and real-time scanning

  • Start the ExamCard.
    The survey and the refscan are performed.

  • Start the 'IViewBOLD' package from the Advanced Viewing Environment (without selecting a
    thumbnail dataset).


  • If the current paradigm is not the correct one, select 'Select paradigm' or 'Edit paradigm' from the 'IViewBOLD' main menu bar.


  • Instruct the patient and manually start the BOLD scan. 


  • View the images including results in real-time with the IViewBOLD package.


Tip 6: About imaging parameters

Synch. external device ('Dyn/Ang' parameter subset)

If set to "Yes", an external device (e.g. laptop) can be synchronized by means of trigger signals being generated at the beginning of each dynamic scan. The FBI box has to be used in this case (see Tip 2).


Start at dyn ('Dyn/Ang' parameter subset) 

This parameter defines the first dynamic at which a trigger pulse will be generated. Negative numbers indicate dummy scans.

Example: With a paradigm of 10 dynamics rest alternating with 10 dynamics stimulation, this parameter has to be set to '11'. In this manner the first new instruction will be presented after the rest period.


Interval (dynamics) ('Dyn/Ang' parameter subset)

This parameter defines the number of dynamic scans between trigger pulses.

Example: With a paradigm of 10 dynamics rest alternating with 10 dynamics stimulation, this parameter has to be set to '10' as the rest and stimulation alternate each 10 dynamics.


Dynamic stabilization ('Dyn/Ang' parameter subset)

This parameter is available for 3.0 T systems. It improves image consistency across dynamic scans and is intended for BOLD Imaging studies utilizing single-shot EPI sequences.

It should be set to YES (applied under all circumstances) or AUTO (applied if applicable).


fMRI echo stabilization

If set to YES, ghost artifacts will be reduced in long-TE FFE scans.


Temporal slice spacing

This parameter controls the timing of slice excitations with a user defined TR. For BOLD imaging, set to 'minimal' so that all slices will be measured directly after each other, thus as much as possible at the same point in time.

Tip 7: About Statistical Parametric Maps

Activated regions of the brain can be visualized with Statistical Parametric Maps.


A Statistical Parametric Map (SPM) is the representation of a statistic in a colored map which is overlayed to the anatomic image. A statistic is a statistical calculation that is performed for each voxel over the dynamics in the functional scan.

ApTip BOLD result SPM bilateral tapping
ApTip BOLD result SPM bilateral tapping

The image shows a finger-tapping Cross-Correlation SPM with the task combinations left versus rest, right versus rest and left versus right.


Types of SPM

Possible types are Cross-Correlation (CC) or Standard Deviation (SD).


The Standard Deviation indicates how much a pixel value is different from the average value in the data set. A small SD indicates that the pixel value is close to the average and that the calculated result in that pixel can be considered as reliable. A large SD will be calculated e.g. if signal changes were caused by movement instead of activation. The SD ranges from 0 to 100.


The Cross-Correlation is a measure of the correspondence of the signal intensity and input paradigm. As such, a paradigm has to be defined in IViewBOLD. A CC of 1.0 is 100% resemblance of the signal curve of the pixel over time and the given paradigm. In a normal BOLD task, CC-values of around 0.35 to 0.45 may be expected when good activation patterns are found.  


With the Cross-Correlation, a task combination has to be defined.


A task combination is defined by two (or more) tasks in two (or more) states that are compared in the statistic.


  • Task A versus Task B
  • Tasks AB versus Tasks CD
  • 1st state versus 2nd state where 1st state stands for 'active' and 2nd state for 'reference'
                                          where 1st state stands for 'left' and 2nd state for 'right'.


A task combination is valid if for each state at least one 'active' dynamic exists while the other state is not. Tasks cannot belong to both states. Dynamics where both states are active are not considered in the computation.



Tip 8: About viewing the SPM's

The anatomic images can be viewed as usual. The overlay of the SPM's can be manipulated by means of the viewing parameters.
ApTip BOLD imaging viewing parameter ikons
ApTip BOLD imaging viewing parameter ikons

The icons in the image above stand for the following viewing parameters: from left to right -

Cluster size  |   Threshold   |      Mask      | Show negative values.


In order to change the parameters cluster size, threshold and mask, click on the up/down arrows of the numerical field besides the icon. In order to show negative values (toggle function), click on the icon.


Cluster size

Cluster (size) is a filter being applied for viewing of the SPMs: only pixels belonging to a minimum sized cluster of the entered size or larger are displayed. A cluster is defined, after application of the threshold, as a group of in-plane pixels that are directly connected. Pixels with positive and negative CC-values are not considered part of one and the same cluster.



For the 'Standard Deviation' statistic, only pixels with a value at or above the threshold value are overlayed.


For the 'Cross-Correlation' statistic, pixels with a value at or above the threshold value are overlayed. If negative values are displayed as well, pixels with a CC-value at or below minus the threshold value will be displayed as well. Thresholds are always equal to or larger than zero.



The mask function is used to avoid SPM calculations for pixels outside of the brain.


Show negative values

If enabled, the negative values are displayed besides the positive values in viewing the SPMs. This option only applies to Cross-Correlation and should be disabled for Standard Deviation.


Tip 9: About the paradigm editor

The paradigm editor is used to setup and edit a paradigm. The information included can be divided into two groups: the tasks and the statistics.



The tasks are to be defined and to be correlated to the corresponding dynamic scans. See more information about tasks and task definition in Tip 1.



Within statistics, it is possible to define the type of SPM to be calculated by default with this paradigm.

ApTip BOLD imaging paradigm editor statistics
ApTip BOLD imaging paradigm editor statistics

Paradigm Editor. Click on |Statistics| to display the window as shown above and to define the parameters.  




Name of SPM

Enter a name for the SPM which is used for further reference to a specified statistic. 


Type of SPM

Either Cross-Correlation or Standard Deviation. See Tip 7 for more information.


Task combination

The task combination is available only with Cross-Correlation. See Tip 7 for more information.


Within the paradigm editor, more parameters which are present via ikons are available.


ApTip BOLD imaging: paradigm editor ikons for statistic
ApTip BOLD imaging: paradigm editor ikons for statistic

Icons from left to right: Smoothing | Delay | Cluster size | Threshold | Color range | Show negative values.


For information on cluster size, threshold and show negative values, refer to Tip 8.


Smoothing (width)

Smoothing is used to prepare images for the statistical calculation and to reduce the noise level. It increases the statistical power at the cost of resolution: the more the images are smoothed (the higher the smoothing width), the higher the lack of resolution.


Smoothing means that the value of every pixel is replaced with a weighted average of a group of pixels around that pixel (also referred to as kernel). This kernel is a square of 1x1, 3x3 up to 9x9 pixels. The pixels in the kernel are weighted according to a Gaussian distribution where the center pixels are more important than the outer pixels.


Smoothing is the Full-Width-Half-Maximum (FWHM) of the Gaussian curve given in pixel sizes. A large width will cause a broad smoothing effect. Note that smoothing is only applied for the kernel. A large width thus requires a large kernel.



The delay parameter accounts for the hemodynamic response. It determines how many dynamics at the beginning of the functional scan are not used for the computation of the statistic. This parameter shifts the task combination by the specified number of dynamics in the calculation of the cross-correlation.


Color Range

The displayed pixels of the SPMs are colored according to gradually changing colors.


Note that the threshold only forms one end of the color range in the paradigm. In the viewers where the SPMs are displayed this can be changed.



color bar pos. neg. values BOLD im.
color bar pos. neg. values BOLD im.

Tip 10: About image quality

As image quality is very much affected by small displacements in between dynamic scans, it is crucial for image quality that the patient lies as quiet as possible.


This content has been made possible by NetForum Community.
Share this on: Share your link in twitter Share your link in facebook Share your link on LinkedIn Print Rate this article: Log in to vote

Apr 15, 2005

Rate this:
Log in to vote

Application Tip
Achieva 1.5T, Achieva 3.0T, Achieva 3.0T TX, Ingenia 1.5T, Ingenia 3.0T, Intera 1.5T, Intera 3.0T
Release 3, Release 5

Clinical News
Best Practices
Case Studies
Publications and Abstracts
White Papers
Web seminars and Presentations
Application Tips and FAQ
Try an Application
Business News
Case Studies
White Papers
Web Seminars and Presentations
Utilization Services
Contributing Professionals
Contributing Institutions
Become a Contributor