A Standards Organization for Open and FAIR Neuroscience: the International Neuroinformatics Coordinating Facility

There is great need for coordination around standards and best practices in neuroscience to support efforts to make neuroscience a data-centric discipline. Major brain initiatives launched around the world are poised to generate huge stores of neuroscience data. At the same time, neuroscience, like many domains in biomedicine, is confronting the issues of transparency, rigor, and reproducibility. Widely used, validated standards and best practices are key to addressing the challenges in both big and small data science, as they are essential for integrating diverse data and for developing a robust, effective, and sustainable infrastructure to support open and reproducible neuroscience. However, developing community standards and gaining their adoption is difficult. The current landscape is characterized both by a lack of robust, validated standards and a plethora of overlapping, underdeveloped, untested and underutilized standards and best practices. The International Neuroinformatics Coordinating Facility (INCF), an independent organization dedicated to promoting data sharing through the coordination of infrastructure and standards, has recently implemented a formal procedure for evaluating and endorsing community standards and best practices in support of the FAIR principles. By formally serving as a standards organization dedicated to open and FAIR neuroscience, INCF helps evaluate, promulgate, and coordinate standards and best practices across neuroscience. Here, we provide an overview of the process and discuss how neuroscience can benefit from having a dedicated standards body.

Markov models for fMRI correlation structure: Is brain functional connectivity small world, or decomposable into networks?

Correlations in the signal observed via functional Magnetic Resonance Imaging (fMRI), are expected to reveal the interactions in the underlying neural populations through hemodynamic response. In particular, they highlight distributed set of mutually correlated regions that correspond to brain networks related to different cognitive functions. Yet graph-theoretical studies of neural connections give a different picture: that of a highly integrated system with small-world properties: local clustering but with short pathways across the complete structure. We examine the conditional independence properties of the fMRI signal, i.e. its Markov structure, to find realistic assumptions on the connectivity structure that are required to explain the observed functional connectivity. In particular we seek a decomposition of the Markov structure into segregated functional networks using decomposable graphs: a set of strongly-connected and partially overlapping cliques. We introduce a new method to efficiently extract such cliques on a large, strongly-connected graph. We compare methods learning different graph structures from functional connectivity by testing the goodness of fit of the model they learn on new data. We find that summarizing the structure as strongly-connected networks can give a good description only for very large and overlapping networks. These results highlight that Markov models are good tools to identify the structure of brain connectivity from fMRI signals, but for this purpose they must reflect the small-world properties of the underlying neural systems.

Direct voxel-based comparison between grey matter hypometabolism and atrophy in Alzheimer's disease

Although the patterns of structural and metabolic brain alterations in Alzheimer's disease are being refined and discrepancies between them are being underlined, the exact relationships between atrophy and hypometabolism are still unclear. In this study, we aimed to provide a direct comparison between grey matter atrophy and hypometabolism in a sample of patients with clinically probable Alzheimer's disease, using a voxel-based method specially designed to statistically compare the two imaging modalities. Eighteen patients with probable Alzheimer's disease of mild severity and 15 healthy aged controls underwent both high-resolution T1 MRI and resting-state (18)FDG-PET. The MRI data sets were handled using optimized VBM. The PET data were coregistered to their corresponding MRI, corrected voxel-wise for partial volume averaging and spatially normalized using the same parameters as those of their corresponding MRI volume. A differential smoothing was applied on the MRI and PET data sets to equalize their effective smoothness and resolution. For each patient, Z-score maps of atrophy and hypometabolism were created by comparing to the controls data set, respectively averaged to provide the profile of hypometabolism and atrophy, and entered in a voxel-by-voxel SPM analysis to assess the statistical differences between hypometabolism and atrophy. The observed patterns of hypometabolism and atrophy were consistent with previous studies. However, the direct comparison revealed marked regional variability in the relationship between hypometabolism and atrophy. Thus, the hypometabolism significantly exceeded atrophy in most altered structures, particularly in the posterior cingulate-precuneus, orbitofrontal, inferior temporo-parietal, parahippocampal, angular and fusiform areas. In contrast, a few hypometabolic structures among which the hippocampus exhibited similar degrees of atrophy and hypometabolism, a profile that significantly differed from the posterior cingulate. Excessive hypometabolism relative to atrophy suggests the intervention of additional hypometabolism-inducing factors, such as disconnection and amyloid deposition, resulting in genuine functional perturbation ahead of actual atrophy and perhaps of pathology as well. Conversely, in the hippocampus, where disconnection processes are also likely to occur, relative synaptic compensatory mechanisms may be taking place, maintaining neuronal activity in the face of structural alterations.

Anatomically informed interpolation of fMRI data on the cortical surface

Analyzing functional magnetic resonance imaging (fMRI) data restricted to the cortical surface is of particular interest for two reasons: (1) to increase detection sensitivity using anatomical constraints and (2) to compare or use fMRI results in the context of source localization from magneto/electro-encephalography (MEEG) data, which requires data to be projected on the same spatial support. Designing an optimal scheme to interpolate fMRI raw data or resulting activation maps on the cortical surface relies on a trade-off between choosing large enough interpolation kernels, because of the distributed nature of the hemodynamic response, and avoiding mixing data issued from different anatomical structures. We propose an original method that automatically adjusts the level of such a trade-off, by defining interpolation kernels around each vertex of the cortical surface using a geodesic Voronoï diagram. This Voronoï-based interpolation method was evaluated using simulated fMRI activation maps, manually generated on an anatomical MRI, and compared with a more standard approach where interpolation kernels were defined as local spheres of radius r=3 or 5 mm. Several validation parameters were considered: the spatial resolution of the simulated activation map, the spatial resolution of the cortical mesh, the level of anatomical/functional data misregistration and the location of the vertices within the gray matter ribbon. Using an activation map at the spatial resolution of standard fMRI data, robustness to misregistration errors was observed for both methods, whereas only the Voronoï-based approach was insensitive to the position of the vertices within the gray matter ribbon.

MR diffusion-based inference of a fiber bundle model from a population of subjects

This paper proposes a method to infer a high level model of the white matter organization from a population of subjects using MR diffusion imaging. This method takes as input for each subject a set of trajectories stemming from any tracking algorithm. Then the inference results from two nested clustering stages. The first clustering converts each individual set of trajectories into a set of bundles supposed to represent large white matter pathways. The second clustering matches these bundles across subjects in order to provide a list of candidates for the bundle model. The method is applied on a population of eleven subjects and leads to the inference of 17 such candidates.

Basal ganglia and supplementary motor area subtend duration perception: an fMRI study

Brain imaging studies on duration perception usually report the activation of a network that includes the frontal and mesiofrontal cortex (supplementary motor area, SMA), parietal cortex, and subcortical areas (basal ganglia, thalamus, and cerebellum). To address the question of the specific involvement of these structures in temporal processing, we contrasted two visual discrimination tasks in which the relevant stimulus dimension was either its intensity or its duration. Eleven adults had to indicate (by pressing one of two keys) whether they thought the duration or the intensity of a light (LED) was equal to (right hand) or different from (left hand) that of a previously presented standard. In a control task, subjects had to press one of the two keys at random. A similar broad network was observed in both the duration-minus-control and intensity-minus-control comparisons. The intensity-minus-duration comparison pointed out activation in areas known to participate in cognitive operations on visual stimuli: right occipital gyrus, fusiform gyri, hippocampus, precuneus, and intraparietal sulcus. In contrast, the duration-minus-intensity comparison indicated activation of a complex network that included the basal ganglia, SMA, ventrolateral prefrontal cortex, inferior parietal cortex, and temporal cortex. These structures form several subnetworks, each possibly in charge of specific time-coding operations in humans. The SMA and basal ganglia may be implicated in the time-keeping mechanism, and the frontal-parietal areas may be involved in the attentional and mnemonic operations required for encoding and retrieving duration information.

A primal sketch of the cortex mean curvature: a morphogenesis based approach to study the variability of the folding patterns

In this paper, we propose a new representation of the cortical surface that may be used to study the cortex folding process and to recover some putative stable anatomical landmarks called sulcal roots usually buried in the depth of adult brains. This representation is a primal sketch derived from a scale space computed for the mean curvature of the cortical surface. This scale-space stems from a diffusion equation geodesic to the cortical surface. The primal sketch is made up of objects defined from mean curvature minima and saddle points. The resulting sketch aims first at highlighting significant elementary cortical folds, second at representing the fold merging process during brain growth. The relevance of the framework is illustrated by the study of central sulcus sulcal roots from antenatal to adult age. Some results are proposed for ten different brains. Some preliminary results are also provided for superior temporal sulcus.

[Bitemporal lobe dysfonction in infantile autism: positron emission tomography study]

PURPOSE

Childhood autism is a severe developmental disorder that impairs the acquisition of some of the most important skills in human life. Progress in understanding the neural basis of childhood autism requires clear and reliable data indicating specific neuroanatomical or neurophysiological abnormalities. The purpose of the present study was to research localized brain dysfunction in autistic children using functional brain imaging.

PATIENTS AND METHODS

Regional cerebral blood flow (rCBF) was measured with positron emission tomography (PET) in 21 primary autistic children and 10 age-matched non autistic children.

RESULTS

A statistical parametric analysis of rCBF images revealed significant bilateral temporal hypoperfusion in the associative auditory cortex (superior temporal gyrus) and in the multimodal cortex (superior temporal sulcus) in the autistic group (p<0.001). In addition, temporal hypoperfusion was detected individually in 77% of autistic children.

CONCLUSION

These findings provide robust evidence of well localized functional abnormalities in autistic children located in the superior temporal lobe. Such localized abnormalities were not detected with the low resolution PET camera,. This study suggests that high resolution PET camera combined with statistical parametric mapping is useful to understand developmental disorders.

In vivo detection of striatal dopamine release during reward: a PET study with [(11)C]raclopride and a single dynamic scan approach

A new simple method is proposed to detect, using PET and [(11)C]raclopride, changes in striatal extracellular dopamine concentration during a rewarded effortful task. This approach aimed to increase the sensitivity in detection of these effects. It requires a single-dynamic PET study and combines the classic kinetic compartmental model with the general linear model of SPM to provide statistical inference on changes in [(11)C]raclopride time-activity curve due to endogenous dopamine release during two short periods of activation. Kinetic simulations predicted that 100% dopamine increase during two 5-min periods starting at 30 and 60 min after the injection can be detected. Moreover the effects of dopamine release on the [(11)C]raclopride time-activity-curve are different from those induced by CBF increase. These simulated curves were used to construct the statistical linear model and to test voxel-by-voxel in healthy subjects the hypothesis that dopamine is released in the ventral striatum during periods of unexpected monetary gains, but not during periods of unexpected monetary loss. The experimental results are in line with the expected results although the amplitude of the effects due to dopamine release is moderate. The advantages and the limits of this method as well as the relevance of the results for dopamine involvement in reward processing are discussed.

The neural system that bridges reward and cognition in humans: an fMRI study

We test the hypothesis that motivational and cognitive processes are linked by a specific neural system to reach maximal efficiency. We studied six normal subjects performing a working memory paradigm (n-back tasks) associated with different levels of monetary reward during an fMRI session. The study showed specific brain activation in relation with changes in both the cognitive loading and the reward associated with task performance. First, the working memory tasks activated a network including the dorsolateral prefrontal cortex [Brodmann area (BA) 9/46] and, in addition, in the lateral frontopolar areas (BA 10), but only in the more demanding condition (3-back task). This result suggests that lateral prefrontal areas are organized in a caudo-rostral continuum in relation with the increase in executive requirement. Second, reward induces an increased activation in the areas already activated by working memory processing and in a supplementary region, the medial frontal pole (BA 10), regardless of the level of cognitive processing. It is postulated that the latter region plays a specific role in monitoring the reward value of ongoing cognitive processes. Third, we detected areas where the signal decreases (ventral-BA 11/47 and subgenual prefrontal cortices) in relation with both the increase of cognitive demand and the reward. The deactivation may represent an emotional gating aimed at inhibiting adverse emotional signals to maximize the level of performance. Taken together, these results suggest a balance between increasing activity in cortical cognitive areas and decreasing activity in the limbic and paralimbic structures during ongoing higher cognitive processing.

Evidence for abnormal cortical functional connectivity during working memory in schizophrenia

OBJECTIVE

Disturbed neuronal interactions may be involved in schizophrenia because it is without clear regional pathology. Aberrant connectivity is further suggested by theoretical formulations and neurochemical and neuroanatomical data. The authors applied to schizophrenia a recently available functional neuroimaging analytic method that permits characterization of cooperative action on the systems level.

METHOD

Thirteen medication-free patients and 13 matched healthy comparison subjects performed a working memory (n-back) task and sensorimotor baseline task during positron emission tomography. "Functional connectivity" patterns, reflecting distributed correlated activity that differed most between groups, were extracted by a canonical variates analysis.

RESULTS

More than half the variance was explained by a single pattern showing inferotemporal, (para-)hippocampal, and cerebellar loadings for patients versus dorsolateral prefrontal and anterior cingulate activity for comparison subjects. Expression of this pattern perfectly separated all patient scans from comparison scans, thus showing promise as a trait marker. This result was validated prospectively by successfully classifying unrelated scans from the same patients and data from a new cohort. An additional 19% of variance corresponded to the pattern activated by the working memory task. Expression of this pattern was more variable in patients during working memory but not the control condition, suggesting inability to sustain a task-adequate neural network, consistent with the disconnection hypothesis.

CONCLUSIONS

Pronounced disruptions of distributed cooperative activity in schizophrenia were found. A pattern showing disturbed frontotemporal interactions showed promise as a trait marker and may be useful for future investigations.

Delayed verbal memory retrieval: a functional MRI study in epileptic patients with structural lesions of the left medial temporal lobe

In a previous functional magnetic resonance imaging (fMRI) study, we suggested that in left medial temporal lobe epilepsy (LTLE) poor verbal episodic memory performances were sustained by abnormal neocortical and mesiotemporal activations. In the present study, we attempted to examine the evolution of these abnormal neocortical and mesiotemporal activations over 24 h. We thus observed the fMRI brain regions activated during the 24-h-delayed retrieval of a word list in the same sample of healthy control subjects and LTLE patients. In control subjects, a similar left occipitotemporofrontal network was activated during both immediate and 24-h-delayed retrieval conditions. In addition, the 24-h-delayed retrieval also activated a larger parietal region and the right hippocampus. This distributed neocortical and mesiotemporal network was very poorly activated during the 24-h-delayed retrieval in LTLE patients, suggesting the inability to reactivate areas that are keys to retrieving stored information.

Human cortical networks for new and familiar sequences of saccades

Visual exploration is organized in sequences of saccadic eye movements that depend on both perceptual and cognitive context. Using functional magnetic resonance imaging, we studied the neural basis of sequential oculomotor behavior and its dependence on different types of memory by analyzing cerebral activity during performance of newly learned and familiar sequences of eye movements. Compared to a resting condition, both types of sequences activated a common fronto-parietal network, including frontal and supplementary eye fields, and several parietal areas. Within this network, newly learned sequences induced stronger activation than familiar sequences, probably reflecting higher attentional demands. In addition, specific regions were recruited for the performance of new sequences, including pre-supplementary eye fields, the precuneus and the caudate nucleus. This indicates that in addition to attentional modulation, novelty of saccadic sequences requires specific cortical resources, probably related to effortful sequence preparation and coordination as well as to spatial working memory. For familiar sequences, recalled from long-term memory, we observed specific right medial temporo-occipital activation in the vicinity of the boundary between the parahippocampal and lingual gyri, as well as an activation site in the parieto-occipital fissure. We conclude that neuronal resources recruited by the gaze system can change with the familiarity of the scanpath to be executed. This study is important to better understand how the brain implements memorized scanpaths for visual exploration and orienting.

Slice acquisition order and blood oxygenation level dependent frequency content: an event-related functional magnetic resonance imaging study

Many event-related functional magnetic resonance imaging paradigms performed so far have been designed to study a limited part of the brain with high temporal resolution. However, event-related paradigms can be exploratory, therefore requiring whole brain scans and so repetition times (TR) of several seconds. For these large TR values, the slice acquisition order may have an important effect on the detection of event-related activation. Indeed, when the scanning is interleaved, the temporal delay between the acquisition of two contiguous slices can reach a few seconds. During this time, the subject is likely to move, and the haemodynamic response will vary significantly. In this case, the interpolation applied between contiguous slices for motion correction induces a temporal smoothing between voxels that are spatially close but temporally sampled a few seconds apart. This should modify the frequency structure of the response and may impair the detection of short events. We, therefore tested the effect of three acquisition schemes (sequential, sequential with gap and interleaved, INT) at two repetition times (TR=3 and 6 s on six and seven subjects, respectively) on activation detection and frequency content in a visual motion event-related paradigm. Unexpectedly, for large TR (6 s), results were found in favour of the INT acquisition scheme (P<0.05). For smaller TR, no strong bias could be found. Generally, intra-subject variability (across acquisition schemes) is found to be much smaller than inter-subject variability, confirming the importance of multi-subjects analyses. Our study also shows that important physiological information is carried by high frequency components that should not be filtered out.

Transient activity in the human calcarine cortex during visual-mental imagery: an event-related fMRI study

Although it is largely accepted that visual-mental imagery and perception draw on many of the same neural structures, the existence and nature of neural processing in the primary visual cortex (or area V1) during visual imagery remains controversial. We tested two general hypotheses: The first was that V1 is activated only when images with many details are formed and used, and the second was that V1 is activated whenever images are formed, even if they are not necessarily used to perform a task. We used event-related functional magnetic resonance imaging (ER-fMRI) to detect and characterize the activity in the calcarine sulcus (which contains the primary visual cortex) during single instances of mental imagery. The results revealed reproducible transient activity in this area whenever participants generated or evaluated a mental image. This transient activity was strongly enhanced when participants evaluated characteristics of objects, whether or not details actually needed to be extracted from the image to perform the task. These results show that visual imagery processing commonly involves the earliest stages of the visual system.

Transient decrease in water diffusion observed in human occipital cortex during visual stimulation

Using MRI, we report the observation of a transient decrease of the apparent diffusion coefficient (ADC) of water in the human brain visual cortex during activation by a black and white 8-Hz-flickering checkerboard. The ADC decrease was small (<1%), but significant and reproducible, and closely followed the time course of the activation paradigm. Based on the known sensitivity of diffusion MRI to cell size in tissues and on optical imaging studies that have revealed changes in the shape of neurons and glial cells during activation, the observed ADC findings have been tentatively ascribed to a transient swelling of cortical cells. These preliminary results suggest a new approach to produce images of brain activation with MRI from signals directly associated with neuronal activation, and not through changes in local blood flow.

Cerebral mechanisms of word masking and unconscious repetition priming

We used functional magnetic resonance imaging (fMRI) and event-related potentials (ERPs) to visualize the cerebral processing of unseen masked words. Within the areas associated with conscious reading, masked words activated left extrastriate, fusiform and precentral areas. Furthermore, masked words reduced the amount of activation evoked by a subsequent conscious presentation of the same word. In the left fusiform gyrus, this repetition suppression phenomenon was independent of whether the prime and target shared the same case, indicating that case-independent information about letter strings was extracted unconsciously. In comparison to an unmasked situation, however, the activation evoked by masked words was drastically reduced and was undetectable in prefrontal and parietal areas, correlating with participants' inability to report the masked words.

Cerebral gray and white matter reductions and clinical correlates in patients with early onset schizophrenia

Few magnetic resonance imaging studies of schizophrenia have investigated brain tissue volumes and their relation to clinical symptoms in patients with an early age at illness onset. The twofold purpose of the study was to investigate both gray and white matter volumes in schizophrenic men with an early age at illness onset, and to determine whether clinical features correlated with tissue volume changes, using an automated voxel-by-voxel image analysis procedure. Twenty male patients with DSM-IV diagnoses of schizophrenia, and an early age at onset (m+/-SD=19+/-2) were compared with 20 age-matched health men. Magnetic resonance (1.5-T) scans were obtained with an Inversion-Recovery prepared fast gradient echo sequence enhancing gray and white matter contrast. Statistical Parametric Mapping was used for image segmentation and comparison. Patients had significant gray matter reductions in medial frontal gyri, left insula, left parahippocampus, and left fusiform gyrus; bilateral white matter reductions in frontal lobes, and increased total cerebrospinal fluid volume were also observed. Negative symptom scores were negatively related to white matter volumes in cingulate regions, and in the right internal capsule. These findings emphasize a pattern of left-hemisphere gray matter abnormalities, and suggest that fronto-paralimbic connectivity may be altered in men with early onset schizophrenia.

The role of dorsolateral prefrontal cortex in the preparation of forthcoming actions: an fMRI study

The dorsolateral prefrontal cortex (DLPFC) plays a key role in working memory (WM). Yet its precise contribution (the storage, manipulation and/or utilization of information for the forthcoming response) remains to be determined. To test the hypothesis that the DLPFC is more involved in the preparation of actions than in the maintenance of information in short-term memory (STM), we undertook a functional magnetic resonance imaging investigation in normal subjects performing two delayed response tasks (matching and reproduction tasks) in a visuospatial task sequence (presentation, delay, response). In the two tasks, the presentation and delay phases were similar, but the expected response was different: in the matching task, subjects had to indicate whether a visuospatial sequence matched the sequence presented before the delay period; in the reproduction task, subjects had to reproduce the sequence and, therefore, to mentally organize their response during the delay. Using a fMRI paradigm focusing on the delay period, we observed a significant DLPFC activation when subjects were required to mentally prepare a sequential action based on the information stored in STM. When subjects had only to maintain a visuospatial stimulus in STM, no DLPFC activation was found. These results suggest that a parietal-premotor network is sufficient to store visuospatial information in STM whereas the DLPFC is involved when it is necessary to mentally prepare a forthcoming sequential action based on the information stored in STM.

Neural substrates for the perception of acutely induced dyspnea

Little is currently known about the brain regions involved in central processing of dyspnea. We performed a functional imaging study with positron emission tomography (PET) to assess brain activation associated with an important component of dyspnea, respiratory discomfort during loaded breathing. We induced respiratory discomfort in eight healthy volunteers by adding external resistive loads during inspiration and expiration. Brain activation was characterized by a significant increase in regional cerebral blood flow (rCBF) (Z score of peak activation > 3.09). As compared with the unloaded control condition, high loaded breathing was associated with neural activation in three distinct brain regions, the right anterior insula, the cerebellar vermis, and the medial pons (respective Z scores = 4.75, 4.44, 4.41). For these brain regions, we further identified a positive correlation between rCBF and the perceived intensity of respiratory discomfort (respective Z scores = 4.45, 4.75, 4.74) as well as between rCBF and the mean amplitude of mouth pressure swings (DeltaPm), the index of the main generating mechanism of the sensation (respective Z scores = 4.67, 4.36, 4.31), suggesting a common activation by these two parameters. Furthermore, we identified an area in the right posterior cingulate cortex where neural activation was specifically associated with perceived intensity of respiratory discomfort that is not related to DeltaPm (Z score = 4.25). Our results suggest that respiratory discomfort related to loaded breathing may be subserved by two distinct neural networks, the first being involved in the concomitant processing of the genesis and perception of respiratory discomfort and the second in the modulation of perceived intensity of the sensation by various factors other than its main generating mechanism, which may include emotional processing.

Detection of fMRI activation using cortical surface mapping

A methodology for fMRI data analysis confined to the cortex, Cortical Surface Mapping (CSM), is presented. CSM retains the flexibility of the General Linear Model based estimation, but the procedures involved are adapted to operate on the cortical surface, while avoiding to resort to explicit flattening. The methodology is tested by means of simulations and application to a real fMRI protocol. The results are compared with those obtained with a standard, volume-oriented approach (SPM), and it is shown that CSM leads to local differences in sensitivity, with generally higher sensitivity for CSM in two of the three subjects studied. The discussion provided is focused on the benefits of the introduction of anatomical information in fMRI data analysis, and the relevance of CSM as a step toward this goal.

Temporal lobe dysfunction in childhood autism: a PET study. Positron emission tomography

OBJECTIVE

The nature of the underlying brain dysfunction of childhood autism, a life-long severe developmental disorder, is not well understood. Although researchers using functional brain imaging have attempted to contribute to this debate, previous studies have failed to report consistent localized neocortical brain dysfunction. The authors reasoned that early methods may have been insensitive to such dysfunction, which may now be detectable with improved technology.

METHOD

To test this hypothesis, regional cerebral blood flow was measured with positron emission tomography (PET) in 21 children with primary autism and in 10 nonautistic children with idiopathic mental retardation. Autistic and comparison groups were similar in average age and developmental quotients. The authors first searched for focal brain dysfunction in the autistic group by using a voxel-based whole brain analysis and then assessed the sensitivity of the method to detect the abnormality in individual children. An extension study was then performed in an additional group of 12 autistic children.

RESULTS

The first autistic group had a highly significant hypoperfusion in both temporal lobes centered in associative auditory and adjacent multimodal cortex, which was detected in 76% of autistic children. Virtually identical results were found in the second autistic group in the extension study.

CONCLUSIONS

PET and voxel-based image analysis revealed a localized dysfunction of the temporal lobes in school-aged children with idiopathic autism. Further studies will clarify the relationships between these temporal abnormalities and the perceptive, cognitive, and emotional developmental abnormalities characteristic of this disorder.

Partially overlapping neural networks for real and imagined hand movements

Neuroimagery findings have shown similar cerebral networks associated with imagination and execution of a movement. On the other hand, neuropsychological studies of parietal-lesioned patients suggest that these networks may be at least partly distinct. In the present study, normal subjects were asked to either imagine or execute auditory-cued hand movements. Compared with rest, imagination and execution showed overlapping networks, including bilateral premotor and parietal areas, basal ganglia and cerebellum. However, direct comparison between the two experimental conditions showed that specific cortico-subcortical areas were more engaged in mental simulation, including bilateral premotor, prefrontal, supplementary motor and left posterior parietal areas, and the caudate nuclei. These results suggest that a specific neuronal substrate is involved in the processing of hand motor representations.

Distinct cortical areas for names of numbers and body parts independent of language and input modality

Some models of word comprehension postulate that the processing of words presented in different modalities and languages ultimately converges toward common cerebral systems associated with semantic-level processing and that the localization of these systems may vary with the category of semantic knowledge being accessed. We used functional magnetic resonance imaging to investigate this hypothesis with two categories of words, numerals, and body parts, for which the existence of distinct category-specific areas is debated in neuropsychology. Across two experiments, one with a blocked design and the other with an event-related design, a reproducible set of left-hemispheric parietal and prefrontal areas showed greater activation during the manipulation of topographical knowledge about body parts and a right-hemispheric parietal network during the manipulation of numerical quantities. These results complement the existing neuropsychological and brain-imaging literature by suggesting that within the extensive network of bilateral parietal regions active during both number and body-part processing, a subset shows category-specific responses independent of the language and modality of presentation.

A cortical region sensitive to auditory spectral motion

The functional architecture of human auditory cortex is still poorly understood compared with that of visual cortex, yet anatomical and electrophysiological studies in non-human primates suggest that the auditory cortex also might be functionally specialized, in a model of parallel and hierarchical organization. In particular, spectral changes such as the formant transitions of speech, or spectral motion (SM) by analogy with visual motion, could be processed in specialized cortical regions. In this study, positron emission tomography (PET) was used to identify which auditory cortical region are involved in SM analysis. We found that a bilateral secondary auditory cortical region, located in the caudal-lateral belt of auditory cortex, was more sensitive to auditory stimuli containing spectral changes than to matched stimuli with a stationary spectral profile. This result suggests that analogies between sensory systems could prove useful in the research into the functional organization of the auditory cortex.

Working memory control in patients with schizophrenia: a PET study during a random number generation task

OBJECTIVE

The authors' goal was to investigate brain regions involved in the deficiency of working memory control processes in patients with schizophrenia.

METHOD

Regional cerebral blood flow was measured with positron emission tomography in eight men with stabilized schizophrenia and eight healthy men while they were performing a graded random number generation task. Twelve scans were made for each subject. Covariations between randomness of responses and regional activation were analyzed.

RESULTS

The pattern of covariation between randomness of responses and activation in the anterior cingulate and superior parietal regions differed between patients and healthy subjects.

CONCLUSIONS

These results suggest a cinguloparietal dysfunction underlying the impairment of working memory control processes during a random number generation task in patients with schizophrenia.

Visual perception of motion and 3-D structure from motion: an fMRI study

Functional magnetic resonance imaging was used to study the cortical bases of 3-D structure perception from visual motion in human. Nine subjects underwent three experiments designed to locate the areas involved in (i) motion processing (random motion versus static dots), (ii) coherent motion processing (expansion/ contraction versus random motion) and (iii) 3-D shape from motion reconstruction (3-D surface oscillating in depth versus random motion). Two control experiments tested the specific influence of speed distribution and surface curvature on the activation results. All stimuli consisted of random dots so that motion parallax was the only cue available for 3-D shape perception. As expected, random motion compared with static dots induced strong activity in areas V1/V2, V5+ and the superior occipital gyrus (SOG; presumptive V3/V3A). V1/V2 and V5+ showed no activity increase when comparing coherent motion (expansion or 3-D surface) with random motion. Conversely, V3/V3A and the dorsal parieto-occipital junction were highlighted in both comparisons and showed gradually increased activity for random motion, coherent motion and a curved surface rotating in depth, which suggests their involvement in the coding of 3-D shape from motion. Also, the ventral aspect of the left occipito-temporal junction was found to be equally responsive to random and coherent motion stimuli, but showed a specific sensitivity to curved 3-D surfaces compared with plane surfaces. As this region is already known to be involved in the coding of static object shape, our results suggest that it might integrate various cues for the perception of 3-D shape.

Episodic memory in left temporal lobe epilepsy: a functional MRI study

Left medial temporal lobe epilepsy (MTLE) is associated with verbal memory impairment usually related to hippocampal damage. We used functional MRI (fMRI) to investigate the patterns of functional activity in healthy volunteers and MTLE patients engaged in verbal episodic memory tasks to look for evidence of a reallocation of verbal memory in epileptic patients. fMRI data were collected from seven MTLE patients with left-sided hippocampal sclerosis and 10 healthy right-handed control subjects on a 3T scanner. Subjects were instructed to learn a list of 17 words (encoding) and then to recall them (retrieval) on successive trials. Healthy volunteers and patients both exhibited bilateral activation (right higher than left) of the parahippocampal gyrus during the retrieval. This effect was more marked in the control subjects. In contrast to the control subjects, patients exhibited consistent and extensive left prefrontal activations in all the memory tasks. These findings show that verbal memory tasks did not involve the same functional patterns in patients and healthy volunteers. This may be interpreted as a dysfunctional response due to the epilepsy and left hippocampal sclerosis, and could reflect the early onset and progressive course of the disease.

Structural group analysis of functional activation maps

We present here a new method for cerebral activation detection over a group of subjects. This method is performed using individual activation maps of any sort. It aims at processing a group analysis while preserving individual information and at overcoming as far as possible limitations of the spatial normalization used to compare different subjects. We designed it such that it provides the individual occurrence of the activations detected at a group level. The localization can then be performed on the individual anatomy of each subject. The analysis starts with a hierarchical multiscale object-based description of each individual map. These descriptions are then compared, rather than comparing the images directly. The analysis is thus performed at an object level instead of voxel by voxel. It is made using a comparison graph, on which a labeling process is performed. The label field on the graph is modeled by a Markov random field, which allows us to introduce high-level rules of interrogation of the data. The process has been evaluated on simulated data and real data from a PET protocol.

Robust smoothness estimation in statistical parametric maps using standardized residuals from the general linear model

The assessment of significant activations in functional imaging using voxel-based methods often relies on results derived from the theory of Gaussian random fields. These results solve the multiple comparison problem and assume that the spatial correlation or smoothness of the data is known or can be estimated. End results (i. e., P values associated with local maxima, clusters, or sets of clusters) critically depend on this assessment, which should be as exact and as reliable as possible. In some earlier implementations of statistical parametric mapping (SPM) (SPM94, SPM95) the smoothness was assessed on Gaussianized t-fields (Gt-f) that are not generally free of physiological signal. This technique has two limitations. First, the estimation is not stable (the variance of the estimator being far from negligible) and, second, physiological signal in the Gt-f will bias the estimation. In this paper, we describe an estimation method that overcomes these drawbacks. The new approach involves estimating the smoothness of standardized residual fields which approximates the smoothness of the component fields of the associated t-field. Knowing the smoothness of these component fields is important because it allows one to compute corrected P values for statistical fields other than the t-field or the Gt-f (e.g., the F-map) and eschews bias due to deviation from the null hypothesis. We validate the method on simulated data and demonstrate it using data from a functional MRI study.

Ambiguous results in functional neuroimaging data analysis due to covariate correlation

In this note we draw attention to a source of potential ambiguity in functional neuroimaging results when data analysis is based on the resolution of a linear model. This ambiguity arises whenever there exists correlation between the model covariates. A single-subject PET activation experiment helps to illustrate to what extent correlation can affect statistical results interpretation, possibly leading to misinterpretation of part of the activation pattern. This note is intended to clarify this point and to suggest the use of a simple and well-known procedure to deal with these situations. In the Appendix, we suggest a convenient mathematical formulation for statistical tests particularly useful in such cases.

Statistical limitations in functional neuroimaging. I. Non-inferential methods and statistical models

Functional neuroimaging (FNI) provides experimental access to the intact living brain making it possible to study higher cognitive functions in humans. In this review and in a companion paper in this issue, we discuss some common methods used to analyse FNI data. The emphasis in both papers is on assumptions and limitations of the methods reviewed. There are several methods available to analyse FNI data indicating that none is optimal for all purposes. In order to make optimal use of the methods available it is important to know the limits of applicability. For the interpretation of FNI results it is also important to take into account the assumptions, approximations and inherent limitations of the methods used. This paper gives a brief overview over some non-inferential descriptive methods and common statistical models used in FNI. Issues relating to the complex problem of model selection are discussed. In general, proper model selection is a necessary prerequisite for the validity of the subsequent statistical inference. The non-inferential section describes methods that, combined with inspection of parameter estimates and other simple measures, can aid in the process of model selection and verification of assumptions. The section on statistical models covers approaches to global normalization and some aspects of univariate, multivariate, and Bayesian models. Finally, approaches to functional connectivity and effective connectivity are discussed. In the companion paper we review issues related to signal detection and statistical inference.

Statistical limitations in functional neuroimaging. II. Signal detection and statistical inference

The field of functional neuroimaging (FNI) methodology has developed into a mature but evolving area of knowledge and its applications have been extensive. A general problem in the analysis of FNI data is finding a signal embedded in noise. This is sometimes called signal detection. Signal detection theory focuses in general on issues relating to the optimization of conditions for separating the signal from noise. When methods from probability theory and mathematical statistics are directly applied in this procedure it is also called statistical inference. In this paper we briefly discuss some aspects of signal detection theory relevant to FNI and, in addition, some common approaches to statistical inference used in FNI. Low-pass filtering in relation to functional-anatomical variability and some effects of filtering on signal detection of interest to FNI are discussed. Also, some general aspects of hypothesis testing and statistical inference are discussed. This includes the need for characterizing the signal in data when the null hypothesis is rejected, the problem of multiple comparisons that is central to FNI data analysis, omnibus tests and some issues related to statistical power in the context of FNI. In turn, random field, scale space, non-parametric and Monte Carlo approaches are reviewed, representing the most common approaches to statistical inference used in FNI. Complementary to these issues an overview and discussion of non-inferential descriptive methods, common statistical models and the problem of model selection is given in a companion paper. In general, model selection is an important prelude to subsequent statistical inference. The emphasis in both papers is on the assumptions and inherent limitations of the methods presented. Most of the methods described here generally serve their purposes well when the inherent assumptions and limitations are taken into account. Significant differences in results between different methods are most apparent in extreme parameter ranges, for example at low effective degrees of freedom or at small spatial autocorrelation. In such situations or in situations when assumptions and approximations are seriously violated it is of central importance to choose the most suitable method in order to obtain valid results.

Characterizing the response of PET and fMRI data using multivariate linear models

This paper presents a new method for characterizing brain responses in both PET and fMRI data. The aim is to capture the correlations between the scans of an experiment and a set of external predictor variables that are thought to affect the scans, such as type, intensity, or shape of stimulus response. Its main feature is a Canonical Variates Analysis (CVA) of the estimated effects of the predictors from a multivariate linear model (MLM). The advantage of this over current methods is that temporal correlations can be incorporated into the model, making the MLM method suitable for fMRI as well as PET data. Moreover, tests for the presence of any correlation, and inference about the number of canonical variates needed to capture that correlation, can be based on standard multivariate statistics, rather than simulations. When applied to an fMRI data set previously analyzed by another CVA method, the MLM method reveals a pattern of responses that is closer to that detected in an earlier non-CVA analysis.

Attention to one or two features in left or right visual field: a positron emission tomography study

In human vision, two features of the same object can be identified concurrently without loss of accuracy. Performance declines, however, when the features belong to different objects in opposite visual fields. We hypothesized that different positron emission tomography activation patterns would reflect these behavioral results. We first delineated an attention network for single discriminations in left or right visual field and then compared this with the activation pattern when subjects divided attention over two features of a single object or over two objects in opposite hemifields. When subjects attended to a single feature, parietal, premotor, and anterior cingulate cortex were activated. These effects were strongest in the right hemisphere and were, remarkably, unaffected by the direction of attention. In contrast, direction of attention affected occipital and frontal activity: right occipital and left lateral frontal activity were higher with attention to the left, whereas right lateral frontal activity was higher with attention to the right. When subjects identified two features of the same object, parietal, premotor, and anterior cingulate activity was enhanced further, predominantly this time in the left hemisphere. Again, there was no direction sensitivity. Direction-sensitive activation of lateral frontal cortex also was increased. Finally, when subjects divided their attention over opposite hemifields, activity in the direction-sensitive occipital and frontal regions fell to a level midway between those seen during exclusively leftward or rightward attention. Thus, the behavioral efficiency with which we attend to multiple features of a single peripheral object is paralleled by enhanced activity in structures generally active during peripheral selective attention as well as in structures that depend on the specific direction of attention, most notably lateral frontal cortex. In addition, in the direction-sensitive regions, dividing attention over hemifields causes a compromise pattern between the extreme levels obtained during unilateral attention.

Grey matter correlates of syndromes in schizophrenia. A semi-automated analysis of structural magnetic resonance images

BACKGROUND

Symptoms in schizophrenia cluster into syndromes, each of which may be associated with a particular pattern of cerebral blood flow. We sought to investigate whether these syndromes are also related to neuroanatomical changes.

METHOD

A semi-automated method was used to examine structural magnetic resonance images in 12 patients with schizophrenia. The relationship between the relative regional grey matter volume and ratings of the syndromes of psychomotor poverty, disorganisation and reality distortion was investigated.

RESULTS

There was a significant negative correlation between psychomotor poverty score and the relative volume of the left ventro-medial prefrontal grey matter, and a significant positive correlation between disorganisation and the relative volumes of the hippocampus, and the parahippocampal/fusiform gyrus bilaterally.

CONCLUSION

The correlation between psychomotor poverty and left prefrontal grey matter volume resembles that previously seen with prefrontal blood flow in the same patient, suggesting that this functional abnormality is related to an underlying anatomical change.

MRI and PET coregistration--a cross validation of statistical parametric mapping and automated image registration

Coregistration of functional PET and T1-weighted MR images is a necessary step for combining functional information from PET images with anatomical information in MR images. Several coregistration algorithms have been published and are used in functional brain imaging studies. In this paper, we present a comparison and cross validation of the two most widely used coregistration routines (Friston et al., 1995, Hum. Brain Map. 2: 165-189; Woods et al., 1993, J. Comput. Assisted Tomogr: 17: 536-546). Several transformations were applied to high-resolution anatomical MR images to generate simulated PET images so that the exact (rigid body) transformations between each MR image and its associated simulated PET images were known. The estimation error of a coregistration in relation to the known transformation allows a comparison of the performance of different coregistration routines. Under the assumption that the simulated PET images embody the salient features of real PET images with respect to coregistration, this study shows that the routines examined reliably solve the MRI to PET coregistration problem.

Combining spatial extent and peak intensity to test for activations in functional imaging

Within the framework of statistical mapping, there are up to now only two tests used to assess the regional significance in functional images. One is based on the magnitude of the foci and tends to detect high intensity signals, while the second is based on the spatial extent of regions defined by a simple thresholding of the statistical map, a test that is more sensitive to extended signals. The aim of this paper is to combine the two tests into a single test that is more sensitive to a wider range of signals. This combined test is based on an analytical approximation of the distribution of these two parameters (size and height) and is applied in the context of statistical maps. The risk of error in noise-only 2D or 3D volumes is assessed under a wide range of experimental conditions obtained by varying both the resolution of the map and the threshold at which clusters are defined. In addition, we have investigated this new test on simulated signals, and applied it to an experimental PET dataset. The experimental risk of error is close to the predicted one, and the overall sensitivity increases when analyzing a volume containing different types of signals.

Detecting activations in PET and fMRI: levels of inference and power

This paper is about detecting activations in statistical parametric maps and considers the relative sensitivity of a nested hierarchy of tests that we have framed in terms of the level of inference (voxel level, cluster level, and set level). These tests are based on the probability of obtaining c, or more, clusters with k, or more, voxels, above a threshold u. This probability has a reasonably simple form and is derived using distributional approximations from the theory of Gaussian fields. The most important contribution of this work is the notion of set-level inference. Set-level inference refers to the statistical inference that the number of clusters comprising an observed activation profile is highly unlikely to have occurred by chance. This inference pertains to the set of activations reaching criteria and represents a new way of assigning P values to distributed effects. Cluster-level inferences are a special case of set-level inferences, which obtain when the number of clusters c = 1. Similarly voxel-level inferences are special cases of cluster-level inferences that result when the cluster can be very small (i.e., k = 0). Using a theoretical power analysis of distributed activations, we observed that set-level inferences are generally more powerful than cluster-level inferences and that cluster-level inferences are generally more powerful than voxel-level inferences. The price paid for this increased sensitivity is reduced localizing power: Voxel-level tests permit individual voxels to be identified as significant, whereas cluster-and set-level inferences only allow clusters or sets of clusters to be so identified. For all levels of inference the spatial size of the underlying signal f (relative to resolution) determines the most powerful thresholds to adopt. For set-level inferences if f is large (e.g., fMRI) then the optimum extent threshold should be greater than the expected number of voxels for each cluster. If f is small (e.g., PET) the extent threshold should be small. We envisage that set-level inferences will find a role in making statistical inferences about distributed activations, particularly in fMRI.

A critique of a new analysis proposed for functional neuroimaging

Methods for analysing functional imaging data have evolved rapidly over the last ten years. Standardized techniques based on formal mathematical and statistical theory and rigorous empirical validation have been proposed to facilitate comparisons of biological results between laboratories. This paper examines an image analysis technique that appears to identify unexpectedly large numbers of activated brain areas for the data collected. It concludes that the method may not adequately control for false positives, rendering interpretation of the functional anatomy difficult.

Cerebral benzodiazepine receptors in hippocampal sclerosis. An objective in vivo analysis

We objectively delineated the extent of abnormalities in central benzodiazepine receptor (cBZR) binding using [11C]flumazenil (FMZ)-PET and high resolution volumetric MRI in 12 patients with mesial temporal lobe epilepsy (mTLE) associated with unilateral hippocampal sclerosis, who underwent presurgical evaluation and subsequent temporal lobe surgery. We compared parametric 3D-images of regional cerebral [11C]FMZ volume of distribution (Vd) obtained from individuals and groups of patients with [11C]FMZ-Vd from a group of 17 healthy controls using statistical parametric mapping (SPM). The [12C]FMZ-Vd was significantly reduced (P < 0.001) in the affected hippocampus in the group of 12 patients with unilateral hippocampal sclerosis. Analysis of individual patients reflected the group finding and there was no significant abnormality of cBZR binding outside the hippocampus in any patient. We conclude that SPM is an useful objective method for detecting changes in regional cerebral [11C]FMZ binding and that in patients with mTLE and unilateral hippocampal sclerosis reduction of cBZR binding is restricted to the area of hippocampal sclerosis with no abnormalities being detected in the temporal neocortex or elsewhere in the neocortex.

Reproducibility of PET activation studies: lessons from a multi-center European experiment. EU concerted action on functional imaging

PET activation studies are performed widely to study human brain function. The question of reproducibility, reliability, and comparability of the results of such experiments has never been addressed on a large scale. Recently, 12 European PET centers performed the same cognitive activation experiment in a European Union funded concerted action. The experiment involved a standardized and validated cross-lingual experimental and control task involving verbal fluency. Each center contributed at least 6 subjects. In total there were 77 subjects and 247 scans in each of the two conditions, giving 494 scans in total. We have analyzed each center's dataset and pooled datasets using statistical parametric mapping. We present results that address the consistency of these analyses, discuss the factors that influence their sensitivity, and comment on a number of related methodological issues. We used a MANOVA to test for center, condition, and centre by condition effects and found a strong condition and center effect and weaker interactions. The main effect determining reproducibility was the overall sensitivity of the experiment, to which the scanner and number of scans contribute in a major way, with a marked advantage for 3D scanners and a large field of view. An important conclusion is that data from different centers can be pooled to improve the reliability of results, which is of particular importance for studies in patients with rare conditions.

Nonlinear regression in parametric activation studies

Parametric study designs can reveal information about the relationship between a study parameter (e.g., word presentation rate) and regional cerebral blood flow (rCBF) in functional imaging. The brain's responses in relation to study parameters might be nonlinear, therefore the (linear) correlation coefficient as often used in the analysis of parametric studies might not be a proper characterization. We present a noninteractive method, which fits nonlinear functions of stimulus or task parameters to rCBF responses, using second-order polynomial expansions. This technique is implemented in the context of the general linear model and statistical parametric mapping. We also consider the usefulness of statistical inferences, based on F fields, about similarities and differences of these nonlinear responses in different groups. This approach is illustrated with a 12-run H215O PET activation study using an auditory paradigm of increasing word presentation rates. A patient who had recovered from severe aphasia and a normal control were studied. We demonstrate the ability of this new technique to identify brain regions where rCBF is closely related to increasing word presentation rate in both subjects without constraining the nature of this relationship and where these nonlinear responses differ.

A voxel-based method for the statistical analysis of gray and white matter density applied to schizophrenia

We describe a novel technique for characterizing regional cerebral gray and white matter differences in structural magnetic resonance images by the application of methods derived from functional imaging. The technique involves automatic scalp-editing of images followed by segmentation, smoothing, and spatial normalization to a symmetrical template brain in stereotactic Talairach space. The basic idea is (i) to convert structural magnetic resonance image data into spatially normalized images of gray (or white) matter density, effected by segmenting the images and smoothing, and then (ii) to use Statistical Parametric Mapping to make inferences about the relationship between gray (or white) matter density and symptoms (or other pathophysiological measures) in a regionally specific fashion. Because the whole brain sum of gray (or white) matter indices is treated as a confound, the analysis reduces to a characterization of relative gray (or white) matter density on a voxel by voxel basis. We suggest that this is a powerful approach to voxel-based statistical anatomy. Using the technique, we constructed maps of the regional cerebral gray and white matter density correlates of syndrome scores (distinct psychotic symptoms) in a group of 15 schizophrenic patients. There was a negative correlation between the score for the reality distortion syndrome and regional gray matter density in the left superior temporal lobe (P = 0.01) and regional white matter density in the corpus callosum (P < 0.001). These abnormalities may be associated with functional changes predisposing to auditory hallucinations and delusions. This method permits the detection of structural differences within the entire brain (as opposed to selected regions of interest) and may be of value in the investigation of structural gray and white matter abnormalities in a variety of brain diseases.

Intersubject variability in functional neuroanatomy of silent verb generation: assessment by a new activation detection algorithm based on amplitude and size information

We present an experimental evaluation of a new algorithm for the detection of activated areas in brain functional maps. The new algorithm, named HMSD, is based on a hierarchical multiscale description of the difference image in terms of connected objects. Size and magnitude of each object are simultaneously tested with respect to a bidimensional frequency distribution derived using Monte-Carlo simulations under the null hypothesis. In the present work. HMSD was applied to the analysis of a silent verb generation PET activation protocol conducted in six right-handed subjects. Applied to single-subject data. HMSD reveals activation located in the left inferior frontal gyrus in three subjects (two in the pars opercularis, one in the pars triangularis), and in the pars opercularis of the right inferior frontal gyrus in one case, the latter being combined to a crossed cerebellar activation. Overall, single-case results were consistent with the analyses of stereotactically averaged data. Despite a 2D implentation. HMSD detection performances of averaged data were better than that obtained with the 2D version of statistical parametric mapping (SPM) and comparable to that of the 3D version of SPM.

Tests for distributed, nonfocal brain activations

Most approaches to detecting changes in functional brain images assume that activations are focal or very localized. However, the brain's response to cognitive of sensorimotor challenge may be spatially or anatomically distributed. In this paper we consider the usefulness of a test based on the mean sum of squares of statistical parametric maps. The performance of this test is evaluated using simulated and real data and is compared to the gamma 2 test, a test of the size of the activated region, and a focal activation test based on the intensity of local maxima. We demonstrate that the mean sum of squares test is more sensitive to nonfocal signals and propose that it could be used to complement approaches that are more sensitive to focal activations.