In vivo Glx measurements from GABA-edited HERMES at 3T are not consistent with those from short-TE PRESS across scanners, regions, and age groups.

Motivation: The reliability of 80ms HERMES for quantification of Glx (glutamate + glutamine) has not been assessed. Goal(s): Investigate the agreement between paired Glx measurements from GABA-edited HERMES (TE = 80ms) and conventional PRESS (TE = 35ms). Approach: HERMES and PRESS were acquired from two voxels (anterior cingulate cortex and thalamus) across 272 participants, 4 scanners, 2 age groups and 2 diagnostic groups. Results: We observe poor agreement between HERMES and PRESS Glx measurements, with consistent systematic and proportional differences between measurements across all scanners, age groups and diagnostic groups. Impact: In a large sample, we show that HERMES (TE = 80ms) Glx estimates are not consistent with paired Glx measurements from short-TE PRESS, highlighting the need for consideration during MRS sequence selection, integrating data across different acquisitions, and data interpretation.