Houston Methodist. Leading Medicine.
Houston Methodist. Leading Medicine

Department of Neurosurgery

The functional Magnetic Resonance Imaging (fMRI) Laboratory

The fMRI lab of the Department of Neurosurgery was established in 2008 with the goals to 1) provide a clinical service to support treatment planning for surgical interventions and 2) to develop new fMRI acquisition and analysis technologies.

Presently, two fMRI installations are available at clinical MRI scanners. The first installation (MRIX Technologies, Inc.) at a GE Excite 3.0 Tesla HDx scanner allows for the evaluation of inpatients and outpatients with paradigms designed following standards of the American Society of Functional Neuroradiology (http://www.asfnr.org/ paradigms.html). Routinely, visually guided motor and language tasks are performed but also custom-designed tasks (visual and/or auditory stimulus) can be implemented. A second installation (Nordic Neurolab, Inc.) at the Siemens Verio 3.0 Tesla scanner at the outpatient center (OPC) exists for evaluating outpatients only with a similar set of paradigms.

A dedicated workstation (Brainlab Inc.) is available to calculate fMRI activation maps for import into the surgical treatment planning system thereby facilitating trajectory planning prior to the surgical intervention (figure 1).

fMRI Lab | Dept of Neurosurgery
Figure 1: Integration of fMRI activation maps into the Brainlab planning station: Raw fMRI image data is transferred via the DICOM protocol, analyzed, co-registered and displayed overlaid onto high-resolution anatomical images. A planned trajectory can be exported for use with the Brainlab system in the OR.

In addition to clinical scans, these installations are also used for fMRI research by several groups here at TMH and in the Texas Medical Center. Successfully conducted studies include the evaluation set shifting in patients with Parkinson's disease [1], studying neural correlates of visual motion processing [2], of novel Spanish word recognition [3], of initiation of willed movement [4] and of processing of music by the brain [5]. A novel approach utilizes graph-theoretical network analysis tools to identify functional highly interconnected brain regions (hubs) to enhance the quality of surgical treatment planning. [6] (figure 2).

fMRI Lab | Dept of Neurosurgery
Figure 2:
Identification of functional distinct brain regions during the execution of a language task by a CPS patient. Left: Graph-network display of all functionally active voxels in the brain reveals separate clusters. Right: Activation in the left inferior frontal gyrus in red overlaid onto anatomical images corresponds to the cluster marked in yellow in the graph network.

Works in progress include implementing resting fMRI for identifying functional subnetworks to complement or to replace task-related fMRI. The advantage of this approach is its applicability in unresponsive patients unable to follow instructions thereby providing treatment-planning information to the surgeon otherwise unobtainable in this patient population(figure 3).


fMRI Lab | Dept of Neurosurgery
Figure 3:
Resting state correlation analysis (bottom row) succeeds in identifying the motor and language fMRI subnetwork (bottom row) as compared to corresponding task-related fMRI activation maps (top row). In addition, the default mode network, active during rest periods, is equally well reproduced (column on the right).


In addition and often complementary to fMRI image data, also diffusion tensor images (DTI) is routinely acquired. DTI image data may prove helpful in surgical treatment planning. For instance, anterior temporal lobectomy may involve Meyer's loop, a bundle of white matter fibers conveying optical information. The knowledge of the exact location of Meyer's loop may help predict potential post-surgical impairment (figure4).

fMRI Lab | Dept of Neurosurgery
Figure 4:
Localization and display of Meyer's loop in the Brainlab surgical planning workstation prior to anterior temporal lobectomy.


  • York MK, Karmonik C, Grossman RG, Wilde E. The relationship between cognitive functioning and whole brain diffusion tensor imaging abnormalities in Parkinson's disease. Neurology. 2008;70(11 Suppl 1):A286
  • Helekar SA, Bishop JM, Karmonik C, Rosenfield DB. Increased selective Activation of Visual Motion-Sensitive Area in Trained Pianists in Response to Musical Gestures, Annual Meeting of the Society for Neuroscience  2012, New Orleans
  • Fernandez V, Allen M, York MK, Wilde EA, Karmonik C, Strutt AM. Neural correlates of a novel Spanish word recognition fMRI memory task. J Int Neuropsychol Soc. 2013;19(Suppl S1).
  • Karmonik C, Dulay M, Verma A, Yen C, Grossman RG. Brain activation in complex partial seizures during switching from a  goal-directed task to a resting state: comparison of fMRI maps to the default mode network. Conf Proc IEEE Eng Med Biol Soc. 2010;2010:5685-8. doi: 10.1109/IEMBS.2010.5627883.
  • Karmonik C, Brandt AK, Fung SH, Grossman RG, Frazier JT. Graph theoretical connectivity Analysis of the Human Brain while Listening to Music with Emotional Attachment: Feasibility Study. EMBC 2013, Annual Meeting, Osaka, Japan
  • Karmonik C, Fung SH, Verma A, Dulay M, Grossman RG. Graph-Theoretical Analysis of fMRI BOLD Activation Maps for Language Localization in Patients with Complex Partial Seizure reveals Small-World Structure of Brain Connectivity Networks: Initial Experience.
  • Wu TC, Wilde EA, Karmonik C, Yallampalli R, Strutt AM, Jankovic J, et al. Frontostriatal changes detected by DTI tractography in Parkinson's disease dementia. J Int Neuropsychol Soc. 2013;19(Suppl S1)
  • Dulay M, Havins W, Agbayani K, Seidl JT, Karmonik C, Xue Z, Verma A, Kawai M, Grossman RG. Brain Imaging Correlates and Executive Impairments in Individuals with Epilepsy and Comorbid Major Depression. Annual Meeting of the Academy of Neurology, Neurology April 24, 2012; 78(Meeting Abstracts 1): PD3.005
  • Karmonik C, Dulay M, Verma A, Grossman RG. Cost function evaluation for the registration of clinical DTI images onto the ICBM DTI81 white matter atlas. Technol Health Care. 2010 Jan;18(2):145-56.
  • Karmonik C. System, Houston Methodist and Computer-Readable Medium For Magnetic Resonance Diffusion Anisotropy Image Processing (Patent granted 2007).