Fast, pseudo‐continuous arterial spin labeling for functional imaging using a two‐coil system

Abstract

A fast, two-coil, pseudo-continuous labeling scheme is presented. This new scheme permits the collection of a multislice subtraction pair in <3 s, depending on the subject's arterial transit times. The method consists of acquiring both control and tag images immediately after a labeling period that matches the arterial transit time. Blood oxygenation level-dependent (BOLD) contrast functional MRI (fMRI) is currently the dominant technique for functional imaging, and it has produced a wealth of information about the brain's cognitive function. Although BOLD techniques offer great detection power, they also have limitations; therefore, there is strong interest in the development of perfusion-based functional imaging techniques. These limitations include the complexity of the mechanism that gives rise to the BOLD signal (1,2), the nonlinearity of that signal (3,4), the temporal noise characteristics (5), and the well-known sensitivity to susceptibility artifacts.In contrast, perfusion is a quantifiable physiological parameter that is easier to relate to neuronal metabolism. Furthermore, our ability to dynamically measure cerebral blood flow (CBF) is crucial for understanding the BOLD effect. Recent animal studies conducted at high field and high spatial resolution indicated that CBF changes are more localized to the parenchyma than the BOLD effect, consistent with the notion that the BOLD effect is weighted toward draining veins (6,7).In the light of these considerations, it is not surprising that there has been extensive work toward the development of rapid, noninvasive cerebral perfusion measurement techniques over the last decade. Among these, arterial spin labeling (ASL) (8 -11), which employs magnetically labeled arterial water as an endogenous tracer, is the most promising. One additional advantage of ASL techniques is the nature of the noise present along the temporal dimension. fMRI time series data contain severe low-frequency, auto-correlated drifts, the amplitude of which is much larger than the BOLD signal itself (12). These lowfrequency drifts are not present in ASL time series, which effectively enables the use of very low-frequency functional paradigms (13). Furthermore, Desmond and Glover (14) recently showed that the power of an fMRI study is largely determined by the intersubject variance of the activation response intensity. In a BOLD experiment, this variance stems from both physiological and hardware variability, whereas in an ASL experiment, the variability observed is mostly due to physiological sources, since the hardware effects are largely subtracted out.ASL techniques also present a number of challenges. These techniques suffer from low SNR, since Ͻ10% of the water in a given voxel is contributed by blood (15,16), and the label decays at a quick rate.

Quantification of perfusion fMRI using a numerical model of arterial spin labeling that accounts for dynamic transit time effects

“…With Turbo-CASL one can collect samples at a much faster rate compared to standard CASL techniques, and still allow for a long labeling time. In our previous publication (1) we showed that the Turbo-CASL method also had an advantage in terms of SNR per unit of time, even though the amount of accumulated tag is not allowed to reach its steady-state maximum. The method is very sensitive to transit time changes.…”

“…We recently presented a new two-coil approach to continuous arterial spin labeling (CASL), (dubbed Turbo-CASL) (1), that significantly enhances the temporal resolution of ASL measurements while the SNR of the measurement is maintained. An additional benefit of the technique is that its sensitivity to transit time changes can be leveraged to increase the sensitivity to brain responses.…”

Towards quantification of blood-flow changes during cognitive task activation using perfusion-based fMRI

“…However, for the primary cortices, such as M1, a selection of data is available. The 26% CBF change in M1 upon pressing the Yes/No button during the Stroop task can be associated with CBF changes in M1 of the order of 60 -100% observed during more demanding finger-tapping tasks (Hernandez-Garcia et al, 2004;Mildner et al, 2003). Although such differences seem reasonable, more data are required to interpret the CBF changes as ''degree of activation''.…”

“…For this purpose, we use data from a recent finger-tapping study (Hernandez-Garcia et al, 2004), which reported an arterial transit-time change of 150 ms and a signal increase of 83%, both averaged over five subjects. The relative blood-flow increase obtained from this signal increase is about 66%.…”

“…This study had utilized a post-label delay (PLD) of the order of the maximum transit time to reduce the transit-time dependence of the CASL signal change (Alsop and Detre, 1996). If short PLDs are employed, functional changes of the transit time additionally contribute to the CASL signal change upon stimulation (Gonzalez-At et al, 2000;Hernandez-Garcia et al, 2004). In the latter study, a dual-coil setup with a sampling interval of the order of the arterial transit time was employed to map CBF changes in the primary motor cortex during finger tapping.…”