1. Supplemental Table S1a: Nutritional composition of standardised meal consumed the evening prior to MRI scans
2. Supplemental Table S1b: Nutritional information of meals consumed during the study day

Supplemental Table S2: Exploratory analysis of endpoints divided by IBS subtype. P value relates to testing for differences between the subtypes using Mann-Whitney tests or unpaired t-tests as appropriate. AUC, area under the curve; IBS-C, constipation-predominant irritable bowel syndrome; IBS-D, diarrhoea-predominant irritable bowel syndrome.

Supplementary Table S3: Area under the curve of colon volumes (in L.min, mean ± SD) for the ascending, transverse, descending and sigmoid colon for each test drink. P value relates to testing for differences using analysis of variance between the test drinks for each region of the colon.

1. a significantly greater than dextrose, p<0.05
2. b significantly greater than dextrose, p<0.005
3. c significantly greater than dextrose, p<0.0005

MRI data analysists were blinded to the intervention received. A range of MRI sequences were used to image the abdomen to obtain the various endpoints including:

1. 1) Colonic gas was assessed as previously published1 using a dual echo gradient echo sequence (TR 175 ms, TE1= 2ms, TE2 = 4.3 ms, FA 80o, ASSET 2) to acquire 24 coronal images with a slice thickness 7mm (no gap) and reconstructed in plane resolution of 1.76 x 1.76 mm2. This sequence was used to measure colonic volumes as well and was acquired during a breathhold. An additional identical sequence was also acquired with the R.F. power set to zero to acquire the noise distribution across the images for gas measurements.
2. 2) Small bowel water content was measured as previously reported2 using a single shot fast spin echo (SSFSE) sequence with fat saturation (TEeff = 325 ms, Echo spacing 5ms) to acquire 32 coronal images with a 7mm slice thickness (no gap) and reconstructed in plane resolution of 0.78 x 0.78 mm2.
3. 3) Colonic volumes were measured from the dual echo images as previously reported3 using MIPAV software4 to segment the different colonic regions.

Gas production from the fermentation of the test substrates was measured using the ANKOM RF gas production system (ANKOM Technology, Macedon, NY, USA). Briefly, per 125ml bottle, 0.5g of psyllium, inulin, or dextrose were added to each fermentation bottle. Additionally, 0.5g of psyllium and 0.5g of inulin were also added to bottles to test the impact of the substrates combined on gas production (inulin + psyllium). To remove any discrepancies in gas produced from sources other than the test substrates, a non-substrate blank was used. To each fermentation bottle, 76ml of media, 5ml of a vitamin and phosphate buffer solution, and 1ml of the reducing solution5 were added under a constant stream of CO2. Once sealed, the substrates were allowed to hydrate, and bottles prewarmed overnight at 37°C.

Bottles were seeded with faecal samples from eight of the IBS individuals from the human MRI study (four IBS-C, four IBS-D). Faecal samples were frozen at -80°C, therefore prior to testing, each faecal sample was defrosted at room temperature. Once defrosted they were diluted in pre-reduced PBS (10% wt/vol), homogenised in a stomacher and strained to remove particulates. Each substrate was fermented in triplicate per volunteer faecal sample. Each bottle was inoculated with 3ml of slurry, sealed, and incubated at 37°C in a shaking water bath (80 rpm) for five days.

Using the ANKOM RF system, the gas pressure was automatically measured every 15 minutes. Gas production from fibre was calculated using previous methods6. Data are reported as cumulative gas volume produced during fermentation, averaged from eight IBS individuals and measured in triplicate per individual/substrate type, thus a total of 24 individual fermentations were performed per substrate.

1. 1. Use the complete kit provided. It contains:
2. 2. If possible, urinate before stool collection to avoid mixing urine and stool as this may affect the sample. THEN PLEASE WASH HANDS YOUR HANDS
3. 3. Place or hold the cardboard tray to collect the stool as passed. THEN PLEASE WASH HANDS YOUR HANDS
4. 4. After opening your bowels, put on the gloves provided (ensure your hands are dry or these will be difficult to put on).
5. 5. Place a large amount of the stool passed into the tube so that it is at least 75% full. To do this, use the spoon attached to the tube as seen in the image below.
6. 6. Place the spoon and stool sample into the tube and lightly secure the lid. DO NOT SCREW THE TOP ON TIGHTLY.
7. 7. After stool collection the remaining stool sample should be flushed away. The rubber gloves and cardboard tray should be placed in a plastic bag and disposed of in your bin as normal rubbish.
8. 8. Place the tube with stool samples into the transparent plastic bag and seal the bag tightly.

1. 9. Place the transparent plastic bag and the ice packs inside the cooler bag/ container provided. This should be sealed and placed in your freezer.
2. 10. Ideally the stool sample should be returned to us as soon as possible (within a few days). Please keep the sample frozen. When you are ready to take the sample to the research site, the cooler bag can be removed from the freezer and taken to your next appointment.

1. Supplemental Figure S3

Water content (ml)

Meal

400

300

200

100

0

Psyllium

Inulin and Psyllium

Inulin

1. Supplemental Figure S3: Small bowel water content (SBWC) rose after test drinks containing psyllium but hardly at all after dextrose. Comparison of areas under the curve (AUCs) showed psyllium was associated with the highest values, significantly greater than inulin plus

1. Supplemental Figure S4

30000

20000

10000

0

Inulin PsylliumInulin + PsylliumCalculated Inulin + PsylliumDextrose

1. Supplemental Figure S4: Tukey box and whiskers plot of area under the curve (ml.hr) after 48 hours of in vitro gas production, divided by IBS subtype (n=4 for each). No significant differences were found between subtypes. IBS-C, constipation predominant irritable bowel syndrome; IBS-D, diarrhoea predominant irritable bowel syndrome.

BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance

Supplemental material placed on this supplemental material which has been supplied by the author(s) Gut

volumes being based on MRI assessment by Pritchard et al3. Analysis was performed with the AR-G2 magnetic bearing rheometer (TA instruments) with a cup (diameter: 30 mm) and vane (bob diameter 28 mm). Inulin was prepared by solubilising in boiling water and storing at 4oC overnight. Psyllium was mixed with the water or the inulin solution immediately prior to analysis. Parameters for data within oscillation strain sweeps ranging from 10-3to 103at 37oC and a frequency of 6.28 rad/s (A, C).

Data shows that the addition of inulin to the psyllium did not alter the viscosity compared to psyllium alone (A). Varying the temperature from room to body temperature did not affect results (B). The 4% solution was more viscous than the 0.5% solution, but both solutions yielded gel structures as demonstrated by an increase in G’ (storage modulus) compared to the G” (loss modulus). These gels also behaved the same whereby they lost their structure, becoming a liquid at the same oscillation strain (C).