This monocentric, prospective observational study was approved by the local ethical commission and followed the tenets of the Declaration of Helsinki. All studied individuals, age matched healthy controls, (n = 52 eyes, age range from 50 to 87 years) and NTG patients (n = 49 eyes, age range from 51 to 84 years), gave their written informed consent.

NTG was diagnosed on the basis of glaucomatous optic disc morphology and concomitant visual field defects with IOP always (without treatment) < 21 mmHg. Glaucomatous changes included optic disc cupping, with or without notches in the neuroretinal rim and localised or segmental loss of retinal nerve fibre layer. Visual field defects were shown by standard automated perimetry (SAP) (Programme G2, Octopus Haag-Streit, Switzerland).

Each patient underwent full ophthalmologic examination including slit lamp assisted biomicroscopy, applanation tonometry, gonioscopy, measurement of central corneal thickness and neuroretinal rim assessment with Heidelberg OCT (HEYEX, red sector on colour code compared to normative database, Heidelberg engineering, CA, USA). None of the patients had a refractive error > +3 or < −3 dioptres (D). 8/26 patients (6 bilateral, 2 unilateral) had IOP lowering drops to keep the IOP as low as possible. 12/26 patients underwent cataract surgery (10 bilateral, 2 unilateral) and none of them had glaucoma surgery. The mean glaucomatous visual field defect (MD) at time of MRI was 12 ± 6 dB and the mean IOP 12 ± 3 mmHg. None of the included NTG patients suffered from any other eye disease or underwent previous posterior segment surgery that may affect the visual field.

All control subjects were volunteers and none of the controls had glaucomatous optic nerve morphology, nor were they on anti-glaucomatous treatment. For each healthy control, the anatomical MRI scan of the ON was normal.

All subjects (NTG and controls) were classified into 4 groups: group I (50–59 years old; nNTG = 10 ONs, mean age: 53 ± 1 years, 6 males; ncontrols = 12 ONs, 54 ± 3 years, 3 females), group II (60–69 years old; nNTG = 15 ONs, mean age: 65 ± 3 years, 6 females; ncontrols = 16 ONs, 64 ± 3 years, 5 females), group III (70–79 years old; nNTG = 18 ONs, mean age: 74 ± 3 years, 7 females; ncontrols = 18 ONs, 75 ± 3 years, 4 females) and group IV ( > 80 years old; nNTG = 6 ONs, mean age: 83 ± 1 years, 3 males; ncontrols = 6 ONs, 86 ± 2 years, 1 female) (Table 1).

The flow velocity between the frontal lobe SAS and the SAS of the ON was calculated and presented in ratios. This was then compared between NTG’s and control’s age-categories.

Images were acquired with a 3 T whole body magnet (Skyra; Siemens Healthcare, Erlangen, Germany) with a 32-channel head coil using Stejskal-Tanner diffusion sequence using following parameters: b = 50 s/mm2, TE/TR = 65/2000 ms, 6 slices, 1 mm slice thickness with acquisition time of 4.13 min, each slice acquired 120 times. Estimating the flow velocities of coherent moving particles through phase contrast images is described in detail in Boye et al. [11]. In-house code programmed in Matlab (MathWorks) was used for image analysis [11]. Shortly, the monopolar diffusion gradients of the diffusion sequence led to a constant phase shift for coherently moving particles. By using the b-values, the maximum encoded velocity (venc) before a phase wrap occurs can be solved and the phase shift can then be used to determine the flow velocity range. Since the phase shift of the diffusion sequence is highly irregular, results are presented as Flow Range Ratio (FRR), which allows the easy comparison between different groups.

Statistical analysis was performed with unpaired t-test performed by SPSS Statistics Software version 21 (IBM Corporation, Armonk, NY, USA) to compare the FRR results between the (i) NTG patients and healthy controls within the age groups, (ii) NTG patients within the age groups as well as (iii) healthy controls within the age groups. For both groups, every second OD and every second OS ON was chosen for statistics.