Methods
The study protocol was approved by the University of Newcastle Australia Human Research Ethics Committee. Informed consent was obtained from all participants who took part in the study in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).

Participants
We present the data from 25 adolescents (mean age 16.8 ± 2.9 years, 10 male subjects) with genetically confirmed 22q11DS recruited from the VCFS & 22q11 Foundation, Australia, and from local health services. The control subjects (HC) were 30 siblings of 22q11DS participants or other typically developing adolescents recruited from the community (mean age 16.5 ± 3.5 years, 14 male subjects). Exclusion criteria for the 22q11DS participants were the presence of the clinical phenotype of 22q11DS without a confirmed 22q11.2 deletion, a clinically detectable medical disorder known to affect brain structure (e.g. epilepsy or hypertension) or a history of head injury. Exclusion criteria for the HC sample were the presence of a genetic disorder or a major mental health problem. Additional exclusion criteria for both groups included a history of severe head injury, seizure disorder or other ocular, neurological or medical problems that could influence task performance. Visual acuity and auditory thresholds were assessed in all subjects prior to testing using a Snellen chart and audiometric assessment of hearing (range: 250 Hz–6 kHz). A psychiatrist (author: US) conducted a structured diagnostic interview [K-SADS; [29]] at the time of testing to determine 22q11DS participants’ diagnostic status. Eleven individuals with 22q11DS were identified as having (or having had) psychiatric diagnoses such as ADHD (n = 3), oppositional defiant disorder (n = 2), generalized anxiety disorder (n = 2), obsessive-compulsive disorder (n = 3), trichotillomania (n = 1) and major depressive disorder (n = 1). Six individuals were currently on medication; antipsychotics: risperidone (n = 2), mood stabilizers: valproate (n = 2), methylphenidate (n = 4) or SSRIs (n = 4). One participant was diagnosed with schizophrenia at the time of interview. The data of this participant was excluded because of non-compliance with task instructions for antisaccade and PPI.a

IQ and executive functioning

Intellectual functioning
The Wechsler Abbreviated Scale of Intelligence [WASI 4 subscale version; [30]] was used to assess general intellectual functioning with verbal, performance and full-scale IQ calculated.

Planning task
A computerised planning task based on the Tower of London (ToL) task was designed. A goal and start configuration was shown simultaneously in the upper and lower half of the screen. The goal configuration was smaller in size and surrounded by a distinct square. An image of three pegs of decreasing heights and three balls of different colours were displayed. The task was displayed on a touch screen. Only one ball could be moved at a time. A ball could only be moved if there was no other ball on top, and three balls could be placed on the long peg, two on the medium and one on the shortest peg. The participants were instructed to transform the start state into the goal state by touching the appropriate ball (to activate it) and then touch the end-position to move the ball. The planning task ranged from 2–7 moves necessary for completion. There was no time limit to solve a problem. Variables recorded were moves above minimum, initial thinking time and subsequent thinking time. All participants received feedback at the end of each trial, consisting of a smiley face in different colours (yellow for perfect performance and purple for passing the trial) or a sad face when failing the trial. All participants also completed a motor control task, in order to control for motor movement time. This consisted of a five-move task that was broken down into 5 one-move tasks so that there should be no thinking time, and all time should be taken up by movement. This time was subtracted from the movement execution time. The participants completed the practice trials before commencing the task and followed by a two-move condition to understand the goal of the main task. These trials were not included in the data analysis.

Experimental tasks

PPI measures
Electromyographic (EMG) recordings of sensorimotor gating of the acoustic startle eyeblink response were undertaken. Auditory stimuli were generated using a presentation software (Neurobehavioural Systems) and were presented binaurally through stereo headphones using a set-up similar to earlier published studies [31]. The startle stimuli were rectangular white noise (50-ms duration, 110 dB sound pressure level (SPL)). Pre-pulse stimuli consisted of two pure tones (high-pitch 1,400 Hz, 20 ms, 5-ms ramp or a low-pitch 800 Hz, 20 ms, 5-ms ramp) at 80 dB SPL (against ~70 dB SPL background noise) presented either 120 or 480 ms before startle stimulus onset. All acoustic stimuli were calibrated using a Bruel and Kaejer sound level meter (type 2231) and artificial ear (type 4151). There were three trial types: (i) startle baseline (startle probe only); (ii) 120-ms pre-pulse and (iii) 480-ms pre-pulse. There were several practice trials for each of the attended and unattended conditions and a total of ten instances in each condition (120 ms passive/active; 480 ms passive/active). The stimulus sequence commenced with the presentation of three startle baseline trials that were discarded and excluded from the analyses. This was followed by an alternating sequence of 20 startle baseline and 20 pre-pulse trials with a variable interstimulus interval of 10–12 s. Within this sequence, 120-ms and 480-ms pre-pulse trials were presented equally in a pseudorandomised order. Two conditions were presented, a passive and an active with the passive condition completed first by all participants. In the passive condition, participants were instructed to ignore the stimuli and make no overt response while watching a silent movie. For the active condition, participants were instructed to listen to the stimuli and were required to respond to the high-pitch tone by pressing a button as quickly and accurately as possible.

EMG recording and analyses
Bipolar silver/silver chloride electrodes were positioned above the orbicularis oculus muscle of the subject’s left eye to record the blink response (A/D rate of 1,000 Hz and amplified × 500). Impedances were reduced to less than 5 kΩ. EMG activity was band-pass filtered (1–200 Hz with 50 Hz notch). Epochs were extracted from 50-ms pre-startle to 300-ms post-startle stimulus onset, baseline corrected over the 50-ms pre-stimulus interval, rectified and averaged separately for each of the three trial types. Startle response amplitude was determined as the integral averaged under the curve occurring between 30 and 150 ms from the onset of startle stimulus. Eight 22q11DS participants did not complete the EMG session. Seventeen individuals with 22q11DS undertook PPI recordings (mean age 16.7 ± 2.8 years, 5 male subjects) and 19 HC (mean age 16.3 ± 4.0 years, 8 male subjects).

Antisaccade recording and analysis
Participants’ eye movements were recorded remotely using an Eyelink 1000 (1,000 Hz, SR Research, Ontario, Canada) linked to a host Dell Pentium IV PC processor and an auxiliary video display unit for observing the monitored eye. Each stimulus block was preceded by a calibration and validation procedure that required participants to fixate on a 3 × 3 matrix of centrally and peripherally located points on the computer screen. Each trial consisted of the following sequence: (i) a circular target was presented at the beginning of each trial; (ii) after 1,000 ms, the target was extinguished, a peripheral target was illuminated and a brief beeping signal was initiated and (iii) an extinction of the cue occurred after 3,000 ms or earlier if participant gaze was recorded within 1° (visual angle) of the target response area and antisaccade target had been initiated for greater than 1,000 ms. The target stimulus, a small red target of 0.9° in diameter, was presented on a black computer screen. The stimuli in each block were presented to the left or right of the screen in a balanced pseudorandom order. An antisaccade error occurred when the subject made a reflexive saccade towards the target before correctly making a saccade in the opposite direction. Antisaccade errors were identified using custom software and expressed as a percentage of the total. Seventeen individuals with 22q11DS undertook antisaccade recordings (mean age 17.1 ± 3.08 years, 7 male subjects) and 28 HC (mean age 16.7 ± 3.4, 13 male subjects).

Analysis
SPSS 19 was employed for statistical analyses. In both participant groups, with the exception of antisaccade error and latency, the distribution of inhibition measures did not violate assumptions of normality (Shapiro–Wilks test) and parametric statistics were used in subsequent analyses.
PPI was determined as a percentage and calculated for each ISI in each stimulus condition using the formula: Percent PPI = 100 × [(startle pulse only units - pre-pulse-startle pulse response units)/startle pulse only units]. PPI measures were examined using repeated measures t-tests and mixed model ANOVA (pre-pulse interval [120 ms, 480 ms] × attention [passive, active] × group). Antisaccade accuracy, amplitude and latency (of correct and incorrect trials separately) were calculated using software designed for eye movement analysis, and antisaccade accuracy data were examined using one-way ANOVAs. Independent samples t-tests, chi square and one-way ANOVA were used to examine demographic and executive functioning differences between groups. Parametric correlational analyses were conducted to examine relationships between executive functions and performance measures on antisaccade and PPI. Threshold set for significance was 0.05.