Evaluating exchange processes in the human brain: magnetization transfer vs adiabatic rotating frame relaxation methods

Introduction Exchange processes of bulk water protons with the protons contained in the macromolecules of tissue can be investigated with magnetization transfer (MT) experiments, which rely on saturating the “solid” pool by placing a preparation pulse (typically a continuous-wave (CW) pulse) several kHz offresonance from water [1]. Since the frequency of the off-resonance pulse does not have a major impact on the MT effect, the macromolecule pool is believed to generate a homogenously broadened line, and therefore the exchange process exploited by MT would occur between spin with identical chemical shift (δω=0, i.e. isochronous exchange IE). By progressively incrementing the duration of the off-resonance pulse, the T1 of water in presence of saturation (T1sat) and the steady state magnetization (Mss) can be estimated, and the exchange rate kf can be extracted as (1/T1sat)(1-Mss/M0). Investigation of exchange processes between spins with either identical or different chemical shifts (δω≠0, anisochronous exchange AE) can be exploited also by rotating frame longitudinal, T1ρ, and transverse, T2ρ, relaxations [2]. A typical method to measure T1ρ and T2ρ is the conventional spinlock experiment, where a CW pulse is applied on-resonance or few hundred Hz off-resonance. Alternatively, rotating frame relaxation experiments can be performed during adiabatic pulses [3], during which the pulse frequency is swept typically over 1-3 kHz around the water resonance. These adiabatic methods have proven to be robust and artifact-free for in vivo applications, and offer the possibility to modulate MR contrast by using different pulse modulation functions [4]. Notably, by proper modeling of the relaxation processes, the simultaneous analysis of T1ρ and T2ρ during adiabatic pulses with different modulation functions allows the extraction of kf for both IE and AE processes [5]. The present work aims at demonstrating different ranges of sensitivity to exchange processes of MT vs adiabatic relaxations experiments in the human brain at 4 T, relying on a quantitative analysis of the relaxation decays and on the formation of the steady state during MT. Methods Six healthy subjects were investigated on 4-T/90-cm Oxford magnet interfaced to Varian INOVA console. Images were acquired using fast spin echo readout, TR = 5 s, TE = 0.60 s, matrix 128 x 128, FOV = 20 cm x 20 cm, and slice-thickness = 3mm. In the adiabatic T1ρ configuration, a train of 4, 8, 12, or 16 HS1 or HS4 pulses was placed prior to the imaging readout, while in the adiabatic T2ρ configuration the train of adiabatic pulses was placed between two 4-ms adiabatic half passage pulses. RF peak power ωmax/(2π) of the adiabatic pulses was 0.88 kHz and 0.625 kHz for HS1 and HS4, respectively. Pulse length was 0.006 s, and the inversion bandwidth was ~1.6 kHz for both HS1 and HS4. For the MT experiment, a 6 kHz off-resonance CW-pulse, with incremental duration (0.2, 0.5, 0.8, 1.0, 1.2 s) and ωmax/(2π) = 0.2 kHz, was placed before the readout. All measurements resulted in similar RF power deposition. The theoretical formalism used to extract exchange parameters during adiabatic pulses have been described previously [5].