Investigation of fMRI activation in the internal capsule

White matter represents approximately half of the tissue in the brain [1]. The idea of white matter activation in fMRI represents an important advance for both basic and clinical studies. Enabling the measurement of functional connectivity more directly than current fMRI approaches, white matter fMRI could provide valuable insight into the dynamics of distributed neural systems and white matter diseases.

However, white matter fMRI is a controversial idea and, up until recently, has been largely disregarded in the literature. The majority of fMRI studies have restricted their focus to gray matter for two reasons; 1) the BOLD signal relies on cerebral blood volume and flow which are 3-6 times lower in white matter [2–6], and 2) the primary source of fMRI signal is thought to arise from post-synaptic potentials as opposed to action potentials [7]. Despite these arguments, white matter tissue has metabolic demands that must be met. This fact has given rise to the idea that white matter activation may be detectable using fMRI. Indeed, an increasing number of fMRI studies have shown white matter activation [8–17].

Most of the white matter fMRI reports have employed tasks that exploit the lateralized nature of the visual and motor systems (opposite response hand to visual hemifield presentation creates a so-called 'crossed' condition) to study interhemispheric transfer and related information processing phenomena. Recently, our group reported prospective white matter fMRI studies at 4T using tasks designed to elicit interhemispheric transfer across the corpus callosum. Mazerolle et al. used a visual Sperry task (word/face) to detect activation in the isthmus of the corpus callosum [13, 18]. These results were observed at the group level and in 20% of the individual subjects (N = 24, p < 0.005, uncorrected). Gawryluk et al. used a Poffenberger task (visual/motor) to elicit activation in the anterior corpus callosum [15, 19]. These results were present at the group level and, notably, in 100% of the individual subjects (N = 10, p < 0.05 corrected).

A key factor that accounted for the sensitivity difference between the two studies related to the method of acquisition. Gawryluk et al. [15] sought to enhance the detection of white matter fMRI using a novel imaging sequence called asymmetric spin echo (ASE) spiral [20]. ASE spiral acquires three images (per slice per volume) with increasing T₂-weighting but equal BOLD-contrast. Sensitivity to white matter fMRI activation increased with increasing T₂-weighting, with the third ASE spiral image (with the highest T₂-weighting) the demonstrating a significant increase in percent signal change relative to the first ASE spiral image (with the lowest T₂-weighting). Moreover, the extent of active voxels in white matter increased as a function of T₂-weighting. The results provided valuable insight into optimizing fMRI acquisition for the detection of white matter activation. Indeed, in a subsequent within-subjects study that administered both the Sperry and Poffenberger tasks using the ASE spiral method of acquisition, corpus callosum activation was observed in 100% and 94% of participants, respectively [17]. These results provide further evidence of the sensitivity of ASE spiral to the detection of white matter fMRI.

There is additional evidence favoring the detection of fMRI signal in white matter. First, white matter activation appears to improve when motion is included as a regressor in the model [13, 17]; this would not be expected if the activation resulted from motion artifact. Second, recent work has shown that white matter activation varies according to task type, indicating that it can be functionally manipulated [17]. Third, as Mazerolle et al. demonstrated, diffusion tensor imaging (DTI) based tractography data can be used to confirm structural connections between active regions in gray and white matter [16].

Taken together, the research to-date has reported white matter activation in the corpus callosum. However, in these previous studies, we cautioned that it is important to verify white matter activity in other structures [13, 15, 17]. This is particularly important if the intent is to develop future applications in both basic science (e.g. the study of functional connectivity) as well as clinical practice (e.g. the assessment of white matter diseases). To extend the current findings, we examined the possibility of detecting fMRI activation in another white matter structure, namely the internal capsule.

To date, the only evidence for white matter fMRI in a fibre tract other than the corpus callosum comes from an abstract by Maldjian et al. [21]. The protocol involved motor tasks with data collected at 1.5 and 4 T. White matter activation was observed in the posterior limb of the internal capsule (PLIC) at 4 T only.

Given that more evidence is needed to characterize controversial white matter activation in pathways outside the corpus callosum, the current study sought to answer the following question: Can white matter fMRI activation be detected in the PLIC using a basic motor paradigm at 4 T? Accordingly, it was hypothesized that it is possible to detect white matter fMRI activation in the PLIC at both the individual and group level.

The current study examined whether white matter fMRI activation could be reliably detected in the internal capsule using a basic motor task and 4 T imaging. As predicted, white matter fMRI activation was detected in the PLIC at both the individual and group level.

These findings correspond with fMRI results obtained by Maldjian et al. [21]. As mentioned, Maldjian et al. collected data at 1.5 and 4 T and observed activation in the PLIC at 4 T only, emphasizing the importance of high field strength to the detection of white matter fMRI [21].

The current findings are also consistent with known neuroanatomy. The PLIC contains corticospinal fibers and is thought to directly connect to the primary motor cortex [22]. Such connections have previously been demonstrated in healthy subjects using a combination of fMRI activation in the primary motor cortex and DTI tractography [23]. Given this, the finding of white matter fMRI activation in the PLIC is thought to be functionally consistent with the motor task employed. The most pronounced activation in the internal capsule was present in the hemisphere contralateral to the engaged hand (Figure 2). This pattern of activation (bilateral: contralateral > ipsilateral) matches that typically seen in the primary motor cortices for a finger tapping task.

One potential concern with imaging a structure as small as the internal capsule is partial volume effects resulting from limitations in resolution. Although the current study used higher resolution than previous work (thinner slices), resolution remains a limitation to be addressed in future research. To address concerns about partial voluming, we examined the individual level data for consistent localization to the PLIC using a double rater approach. Using this method, white matter activation in the PLIC was present in 80% of participants (Figure 1). This proportion of subjects with white matter activation is highly consistent with previous results [15, 17]. This is likely due to the use of ASE spiral imaging in combination with high field MRI.

Interestingly, of participants with activation in the PLIC, 87.5% showed activation in the left hemisphere during right finger tapping and 100% showed activation in the right hemisphere during left finger tapping. Additionally, at the group level, PLiC activation was only observed during left finger tapping. Given that all participants were right handed, this difference may reflect task demand related to hand dominance.

A common question that has arisen is: why is activity not visible along the entire white matter tract? There are at least three key considerations related to this issue. First, it is important to emphasize that the objective of this work is to demonstrate functional MRI sensitivity to activation in white matter, which was previously thought to be undetectable. Given this, the initial evidence is expected to occur in the regions that show the strongest activation. Once sensitivity to the phenomenon is better understood, better visualization is expected. Second, related to point one, the most active regions along a tract can vary. In the current study, individuals showed variability in PLIC activation (Figure 1; Table 1). Evidence from studies on the vasculature supplying the PLIC corroborates these findings. Specifically Donzetli et. al. documented variability in vasculature feeding the region of the PLIC [24]. Third, if WM activity could be explained solely on the basis of false positives, when averaged together as a group, one would not expect a group representation of the white matter structure. Our data show clear a group effect, which in fact is an improved representation of the underlying anatomy (Figure 2). In order to further address the question, we are currently studying white matter activation in an animal model using a combination of cerebral blood flow/volume measurements and electrophysiological techniques.

Despite increasing evidence, the detection of fMRI in white matter remains controversial. However, studies using fluorodeoxyglucose autoradiography in rats have detected activity-dependent metabolic changes in white matter [25]. There is also evidence that spiking activity is correlated with fMRI activation in addition to local field potentials (albeit to a lesser extent) [26–28]. One contributing neurophysiologic source of fMRI signal changes in white matter is increased activity of ATP-dependent Na+/K+ pumps, required to restore ionic gradients that are disrupted by axonal conduction [29–31]. Further investigation of the physiological basis and possible imaging mechanism underlying this phenomenon are ongoing.

Despite the fact that white matter comprises half of the brain, few studies have attempted to measure fMRI activity in this tissue. Recently, we reported fMRI activation in the corpus callosum [13, 15–17]. The current study provides evidence of white matter fMRI activation in the internal capsule. These results represent an important avenue to advance in studies of functional connectivity as well as the clinical assessment of white matter disease/disorder.