Previous fMRS studies conducted at 7T have generally utilized long TR and short TE in order to minimize relaxations effects, while maximizing the sensitivity to changes in metabolite concentrations. Minute changes in the order of 0.2 mmol/g have been thus reliably measured for lactate, aspartate, glucose and glutamate in the human visual cortex during visual activation (1,2). Since at longer TE and/or shorter TR the signal intensities are particularly sensitive to relaxation, it is possible that in such experimental conditions the fMRS results might be modulated by changes in metabolite relaxation times (especially T2) during activation and/or by different relaxation properties of metabolites in different compartments, such as extra-, intracellular lactate. In fact, other studies conducted at lower fields and longer TE (3) reported metabolite changes somewhat different from those reported at ultra-short TE. Therefore, the goal of the present study was to determine functional changes of metabolite signals at long TE and short TR (i.e., when the system is not fully relaxed) during prolonged visual stimulation at 7T.13 healthy young volunteers (age 31 ± 10 years) underwent fMRS sessions at 7T . Metabolite spectra (semi-LASER, TR = 2.5s,TE = 135ms) were acquired during 25-min-long functional spectroscopy paradigm (5min REST, 10 min STIM, 10 min REST. 64-scans blocks of metabolite spectra were quantified in LCModel with basis set that consisted of 22 metabolites and measured spectrum of macromolecules (N=5, TR = 2.5s, TI = 750 ms, Fig 1). For NAA, Cr and phosphocreatine (PCr), CH3 and CH2 resonances were fitted separately in order to accommodate the different relaxation time constants for CH3 and CH2 resonances. Unsuppressed water spectra were acquired during short visual stimulation (30s REST, 30 s STIM) in order to quantify the BOLD effect as we did previously (Bednarik 2015). The visual stimulus consisted of black and white checkerboard flickering at 8 Hz. We applied standard processing to the NMR spectra, created linewidth-matched spectra between REST and STIM in order to take into account the BOLD effect, and finally quantified the resulting spectra with LCModel. Statistical significance of changes in metabolite signal intensities during activation was inferred by two-tailed paired t-test.Water line-widths were in average 13.77±0.89 Hz. The average BOLD effect measured as line-width change on water peaks (~0.5 Hz) was consistent with our previous studies (Bednarik). Glutamate and lactate were quantified with CRLB ~3% and ~16%. Concentration differences (STIM-REST) were calculated between time-points highlighted in Fig. 2 after linewidth matching of the STIM and REST spectra (Bednarik 2015). In response to visual stimulation, Glu and Lac significantly increased by 5%±1%(p = 0.002) and 37%±6% (mean±SEM, p = 0.0005), respectively. These functional changes are slightly higher but consistent with those we reported previously for Glu (3%±1%) and Lac (30%±7%), indicating that the contribution of relaxation during stimulation is not critical for those metabolites. The LCModel results were confirmed by the average difference spectrum (Fig. 3). Small residues in the difference spectrum were also observed even after line-matching at the frequency of the singlets of NAA and total Cr (2.01 and 3.03 ppm, respectively), which might be ascribed to possible small changes of T1 and/or T2 of these metabolites during activation. Long TE fMRS is able to reliably detect changes in metabolite signals which are consistent with those reported at ultra-short TE. Observed small increases in NAA and tCr signals can be ascribed to changes in relaxation of such metabolites during neuronal activation.