Article Figures & Data
Figures
- Figure 1.
Task design. a, Subjects were required to pick one of four options on each trial. Three seconds after selection, the option yield and outcome that included possible reward (20p) or punishment (painful shock) was given. b, Exemplar graph for one option indicating that the probability of reward and punishment was independent and varied slowly over trials. All options were independent of each other. Outcomes were delivered simultaneously and followed by a variable intertrial interval of 2–6 s.
- Figure 2.
a, b, Pre-ingestion (a) and 5 h post-ingestion (b) of serum Tryp:LNAA in the TRP+ (Control) and TRP− (Depleted) groups. c, Current—VAS responses showing a statistically fit sigmoid response function and estimated current intensity for VAS = 8 in an example subject.
- Figure 3.
Behavioral results. a, Probability (frequency/total number of trials) of repeating a choice on a particular option given any outcome (All) and given each of the four possible outcomes: reward alone (Rew), pain alone (Pain), reward and pain (Both), and neither reward nor pain (Neither). b, Independent decision weights in TRP+ and TRP− groups. This shows the weights by which rewards, punishments, or previous choices governed the tendency to repeat the same option. Subjects were on average more sensitive to rewards (20p) than punishments (painful shock). c, Decision weights parameterised as exponential kernels, which is equivalent to a truncated reinforcement learning model. Parameter estimates were TRP+ reward sensitivity = 4.18, punishment sensitivity = −2.29, choice repetition (perseveration) = 0.79, TRP− reward sensitivity = 2.50, punishment sensitivity = −2.43, choice repetition = 1.90. d, e, Correlation between change in serum tryptophan:large neutral amino acid ratio at 5 h post-amino acid ingestion and the parameters for reward sensitivity (r = 0.50, p < 0.005; d) and choice perseveration (r = −0.41, p < 0.01; e).
- Figure 4.
Neuroimaging results: choice values and choice perseveration. a, BOLD responses corresponding to reward value and avoidance value (thresholded at p < 0.01 × p < 0.01 for display) in ventromedial prefrontal cortex 2, 38, −12 (circled)—our region of a priori interest. Note that both appetitive and avoidance values individually satisfy correction for multiple comparisons in this region (FWE ROI p < 0.05, see Materials and Methods). b, BOLD responses corresponding to covariation of reward value and change in serum tryptophan: long chain amino acid ratio following amino acid ingestion, threshold p < 0.005 for illustration (FWE ROI p < 0.05), showing response in ventromedial prefrontal cortex 12 44 −8 (z = 3.52). c, BOLD responses corresponding to covariation of choice kernel and change in serum tryptophan: long chain amino acid ratio following amino acid ingestion, threshold p < 0.005 for illustration, showing responses in right head of caudate (18, 4, 14; FWE p < 0.05 whole brain) and right anterior pole (16, 66, 14).
- Figure 5.
Neuroimaging results: prediction error. a, Overlapping reward and punishment avoidance prediction errors (exclusively masked, p < 0.01 × p < 0.01 for illustration) showing activity in right medial head of caudate (10, 12, 6) and right dorsolateral putamen (28, 0, 6). The appetitive and avoidance prediction error responses are individually significant at FWE ROI p < 0.05 b, Reward prediction errors modulated by serum ΔTRP:ΣLNAA (covariate regression, p < 0.001 uncorrected for display; 34, −4, 4; FWE ROI p < 0.05).
Tables
Isoleucine 4.2 g Leucine 6.6 g Lysine 4.8 g Methionine 1.5 g Phenylalanine 6.6 g Threonine 3.0 g Valine 4.8 g Tryptophan or placebo 3 or 0 g Alert/Drowsy Calm/Excited Strong/Feeble Clear-Headed/Muzzy Well-coordinated/Clumsy Energetic/Lethargic Contented/Discontented Tranquil/Troubled Quick-witted/Mentally slow Relaxed/Tense Attentive/Dreamy Proficient/Incompetent Happy/Sad Amicable/Antagonistic Interested/Bored Gregarious/Withdrawn Reward value p < 0.001 (uncorrected) Medial prefrontal cortex −4, 54, −8 (z = 3.74); FWE p < 0.05 SVC Bilateral head of caudate 14, 2, 16 (z = 4.12); −8, 0, 14 (z = 4.05) Anterior cingulate cortex 2, 44, 12 (z = 3.72) Right lateral orbitofrontal cortex 44, 54, −8 (z = 4.35) Avoidance value p < 0.001 (uncorrected) Medial prefrontal cortex 4, 36, −8 (z = 3.40); FWE p < 0.05 SVC Nucleus accumbens −4, 4, −10 (z = 4.56) Secondary somatosensory cortex/posterior insula −52, 20, 12 (z = 5.45); −38, −12, 16 (z = 5.16) Precuneus −14, −60, 18 (z = 4.97) Posterior parietal cortex 0, −26, 52 (z = 4.66) Overlapping avoidance and reward value Ventromedial prefrontal cortex 2, 38, −12; −6, 52, −8; −6, 28, −12 Dorsomedial prefrontal cortex 2, 60, 24; 16, 58, 34; −18, 60, 22; −16, 66, 14 Left posterior insula −56, −2, 0 Bilateral head of caudate 12, 2, 16; −8, 2, 16 Medial parietal cortex 2, −20, 44 Precuneus −8, −46, 34 Reward × serotonin (ΔTRP:ΣLNAA) p < 0.001 (uncorrected) Ventromedial prefrontal cortex 12, 44, −8 (z = 3.52); FWE p < 0.05 SVC Left anterior insula −54, 16, 6 (z = 3.84) Precuneus 16, −54, 28 (z = 3.23) Reward prediction error p < 0.001 (uncorrected) Bilateral head of caudate 12, 0, 14 (z = 4.11); −4, −4, 12 (z = 4.07); FWE p < 0.05 SVC Cerebellum 20, −52, −24 (z = 5.10) Dorsomedial prefrontal cortex 12, 64, 26 (z = 4.24) Anterior cingulate cortex −4, 36, 28 (z = 3.45) Choice kernel p < 0.001 (uncorrected) Nucleus accumbens −4, 10, −8 (z = 4.29) Dosomedial prefrontal cortex −6, 60, 34 (z = 4.32) Ventromedial prefrontal cortex −12, 44, −4 (z = 3.88) Choice kernal x -serotonin (ΔTRP:ΣLNAA) p < 0.001 (uncorrected) Right head of caudate 18, 4, 14 (z = 4.91); FWE p < 0.05 (whole brain) Anterior cingulate cortex −2, 42, 16 (z = 3.32) Anterior pole 16, 66, 10 (z = 3.37)* Avoidance prediction error p < 0.001 (uncorrected) Right head of caudate 12, 2, 18 (z = 4.08); FWE p < 0.05 SVC Bilateral dorsolateral putamen 30, 0, 8 (z = 3.76); −22, 0, 8 (z = 3.53); FWE p < 0.05 SVC Ventrolateral putamen 30, −8, −6 (z = 4.63) Bilateral amygdaloid complex 24, 2, −12 (z = 3.84); −28, −4, −14 (z = 3.54) Overlapping reward and avoidance prediction error p < 0.01 × p < 0.01 (uncorrected) Right head of caudate 10, 12, 6 Right dorsolateral putamen 28, 0, 6 Reward prediction error × serotonin (ΔTRP:ΣLNAA) p < 0.001 (uncorrected); FWE p < 0.05 SVC Right dorsolateral putamen 34, −4, 4 (z = 4.22) Results in bold are a priori ROIs (ventromedial PFC, choice values; dorsal striatum, instrumental prediction errors; see Materials and Methods). SVC, Small volume correction.
↵*Not SVC as distinct from previous anterior pole activity related to exploration.
