PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • Twenty-seven typically developing children (ages 4.3 to 10.8 y, mean age = 7.1 y, SD = 1.6, 16 female) and 20 adults (ages 18.9 to 25.4 y, mean age = 20.7 y, SD = 1.7, 13 female) successfully participated in one or more of the experimental conditions (26 children and 20 adults in the natural viewing fMRI paradigm, 23 children and 20 adults in the traditional fMRI paradigm, and 19 children in the behavioral standardized testing). Children were excluded from conditions due to excessive head motion (>5 mm), opting-out, or experimenter error. The mean motion deviations for the remaining children (after online motion correction) were 0.39 mm translation (σ = 0.37) and 0.36 degrees rotation (σ = 0.24) in the natural viewing paradigm and 1.26 mm translation (σ = 1.33) and 1.5 degrees rotation (σ = 1.38) in the traditional paradigm. There was significantly less child motion in the natural viewing paradigm compared to the traditional paradigm (translation: t(21) = 3.5, p<0.005; rotation: t(21) = 3.97, p<0.005). The difference in child head motion between tasks is noteworthy considering that the natural viewing task was almost twice as long as the traditional task. Anecdotally, children seemed calmer and more engaged by the natural viewing task than the traditional task and this observation is empirically supported by the motion data. All participants were screened for neurological abnormalities. All procedures were approved by the Research Subjects Review Board. Prior to the MR scanning session, children were given a 30-min training session in a mock scanner to practice the experimental task, and remaining motionless during scanning. In the actual MR scanner, headphones, foam padding, and medical tape were used to secure the children's heads. Adults received verbal instructions and a brief session of task practice. During the MR scanning session, we measured participants' neural activity (BOLD) during (1) a natural viewing paradigm and (2) a traditional fMRI paradigm with faces, shapes, numbers, and words. In the natural viewing paradigm, participants viewed a single 20.3-min montage of clips from children's educational television shows (Figure 1A). Individual clips ranged from 12 to 176 s in length and were edited into one continuous movie. The content of the video included letters, numbers, and other subjects (e.g., planets, shapes, Egypt). Participants were instructed to remain motionless while watching the movie but were given no instructions to fixate or restrict eye movement. A short quiz was administered at the end of the scanning session to ensure that participants attended to the movie. In the traditional fMRI paradigm, subjects compared pairs of stimuli presented on a computer monitor and reported whether the stimuli were the same or different. The stimuli consisted of pairs of isolated images (faces, numbers, words, or shapes) presented to the left and right of a central crosshair. Participants were instructed to fixate on the central crosshair throughout the scanning session. When a pair of stimuli was presented, participants were instructed to press a response button only if the two stimuli matched. No button press was taken to indicate a “non-match”. Fifty percent of the trials, distributed at random across each stimuli type, were “matches,” while the other 50% were “non-matches.” In the “faces” condition, one face image was presented as a frontal shot; the other was presented as an oblique view. The faces were a “match” when both images were of the same person. On “numbers” trials, one image was an Arabic numeral between 1 and 9, and the other was a dot array. The images were a “match” when the number of dots in the dot array matched the Arabic numeral. The “words” condition consisted of trials in which two word images were presented, one in all capital letters in a serif font, and the other in all lowercase letters in a sans-serif font. The words were a “match” when both were the same word. On “shape” trials, two shape images were presented. A “match” occurred when both shapes were identical. Stimuli were presented in 4.4-min runs in a blocked design. Each run consisted of three blocks per condition, with three picture comparison trials from the same condition per block. Each trial was presented for 2 s, followed by a 2-s inter-trial interval. 8 s of fixation followed each block. Blocks were semi-randomly presented. Children were tested on two standardized IQ tests after the scanning session: the Test of Early Mathematics Ability, 3rd Edition TEMA-3 [28] and Kauffman Brief Intelligence Test, 2nd Edition KBIT-2 [29]. Whole brain BOLD imaging was conducted on a 3-Tesla Siemens MAGNETOM Trio scanner with a 12-channel head coil at the Rochester Center for Brain Imaging. High-resolution structural T1 contrast images were acquired using a magnetization prepared rapid gradient echo (MP-RAGE) pulse sequence at the start of each session (TR = 2,530 ms, TE = 3.44 ms flip angle = 7 degrees, FOV = 256 mm, matrix = 256×256, 160 or 176 [depending on head size] 1×1×1 mm sagittal left-to-right slices). An echo-planar imaging pulse sequence with online motion correction was used for T2* contrast (TR = 2000 ms, TE = 30 ms, flip angle = 90 degrees, FOV = 256 mm, matrix 64×64, 30 sagittal left-to-right slices, voxel size = 4×4×4 mm). The first six TRs of each run were discarded to allow for signal equilibration. The “movie” run of the natural viewing paradigm was one functional run of 610 volumes. The traditional fMRI paradigm was distributed over two to four functional runs of 132 volumes each. Total scanning time was approximately 40 min. fMRI data were analyzed with the BrainVoyager 2.1 software package and in-house scripts drawing on the BVQX toolbox in MATLAB. Preprocessing of the functional data included, in the following order, slice scan time correction (sinc interpolation), motion correction with respect to the first (remaining) volume in the run, and linear trend removal in the temporal domain (cutoff: two cycles within the run). Functional data were then registered (after contrast inversion of the first remaining volume) to high-resolution de-skulled anatomy on a participant-by-participant basis in native space. For each individual participant, echo-planar and anatomical volumes were transformed into standardized space [30]. Data from adults and children were normalized into the same Talairach space. The functional data from the traditional fMRI paradigm were not smoothed. A Gaussian spatial filter with an 8 mm full-width at half-maximum was applied to each volume for the natural viewing paradigm. We spatially smoothed the natural fMRI data because of the precedent set by Hasson and colleagues [2] for inter-subject correlations; we used the more conservative smoothing kernel (8 mm) of the two kernels tested in that prior study (8 mm and 12 mm). Functional data from the traditional fMRI paradigm were analyzed using the general linear model (random effects analysis). Experimental events (duration = 10 s) in the traditional fMRI paradigm were convolved with a standard dual gamma hemodynamic response function. There were four regressors of interest (corresponding to the four stimulus types), one regressor for the button press, and six regressors of no interest, corresponding to the motion parameters obtained during preprocessing. For the natural viewing fMRI paradigm, the data were pre-processed as described above for the traditional paradigm, and the resulting timecourses formed the basis for the intersubject correlation analyses. FD [31] was regressed out of each subject's timecourse to control for frame-to-frame head motion. FD is calculated by summing the absolute values of the derivatives from the six motion estimates of translation and rotation. Rotational displacements are converted to millimeters by projecting radians onto a sphere with a 50 mm radius (following Power et al. [31]). Subsequent analyses were performed on the residual timecourses after FD was regressed out. An additional control for signal intensity changes (DVARS following Power et al. [31]) is presented in Figure S2. That method removes volumes (1 back and 2 forward) surrounding timepoints where signal intensity changes by 0.5% or greater. We implemented a developmental intersubject correlation method by correlating the timecourse of each voxel in the brain (for the whole 20-min video) for each child with the corresponding voxel in each adult (paired r-maps). In these children-to-adults correlations, we correlated each subject with every other subject (rather than correlating each child's data with an adult average) in order to be able to carry out parallel analyses for adults-to-adults and children-to-children without including the subjects' own data in the average. Additionally, this approach of correlating each subject with every other subject preserves the variability from individual subjects. After obtaining paired r-maps for each child paired with each adult, we then calculated a mean image across the paired r-maps (a mean r-map) for each child. Each mean r-map represented that child's average similarity to adults. Each child thus had one r-map representing their mean similarity in neural activity to a group of adults at every voxel in the brain. These maps provide an index of how “adult-like” or, mature, each child's neural responses are across the brain, and are referred to as “neural maturity maps.” We performed group-level statistics (one sample t-test on Fisher-transformed r values) over the children's “neural maturity maps” to plot the average group-level similarity of children's natural viewing BOLD timecourses to those of adults (whole brain). That analysis is shown in Figure 1. The same intersubject correlation method was used within-groups for the children-to-children and adults-to-adults correlation maps shown in the right two panels of Figure 1. A whole brain analysis was conducted to measure the correlation between the children's chronological ages and their neural maturity maps (i.e., one correlation per voxel, between the vector of children's ages and their neural maturity values). Similarly, whole brain partial correlations between behavioral tests (TEMA, KBIT) and neural maturity were conducted over the children's neural maturity maps. The whole-brain partial correlations were conducted by regressing one test score out of the other and then correlating the residuals with neural maturity for each voxel, across the subject group. In addition, we also conducted ROI analyses. The regions tested in all ROI analyses were defined with independent data from their statistical tests. Note that some additional ROI data are shown from whole-brain analysis to illustrate individual subject scores. Percent signal change was calculated to illustrate the timecourses from the natural viewing paradigm. Percent signal change was calculated for each subject on the raw timecourse data by dividing each timepoint's intensity value by the mean intensity of the whole timecourse, then multiplying by 100 and subtracting 100. Statistical tests over response amplitudes from ROIs were conducted on the residual timecourses after the FD regression.
rdf:type