A heuristic for the degrees of freedom of statistics based on multiple variance parameters

Crossmodal emotional integration in major depression

Major depression goes along with affective and social-cognitive deficits. Most research on affective deficits in depression has, however, only focused on unimodal emotion processing, whereas in daily life, emotional perception is often highly dependent on the evaluation of multimodal inputs. We thus investigated emotional audiovisual integration in patients with depression and healthy subjects. Subjects rated the expression of happy, neutral and fearful faces while concurrently being exposed to emotional or neutral sounds. Results demonstrated group differences in left inferior frontal gyrus and inferior parietal cortex when comparing incongruent to congruent happy facial conditions, mainly due to a failure of patients to deactivate these regions in response to congruent stimulus pairs. Moreover, healthy subjects decreased activation in right posterior superior temporal gyrus/sulcus and midcingulate cortex when an emotional stimulus was paired with a neutral rather than another emotional one. In contrast, patients did not show such deactivation when neutral stimuli were integrated. These results demonstrate aberrant neural response in audiovisual processing in depression, indicated by failure to deactivate regions involved in inhibition and salience processing when congruent and neutral audiovisual stimuli pairs are integrated, providing a possible mechanism of constant arousal and readiness to act in this patient group.

Shifted neuronal balance during stimulus-response integration in schizophrenia: an fMRI study

Schizophrenia is characterized by marked deficits in executive and psychomotor functions, as demonstrated for goal-directed actions in the antisaccade task. Recent studies, however, suggest that this deficit represents only one manifestation of a general deficit in stimulus-response integration and volitional initiation of motor responses. We here used functional magnetic resonance imaging to investigate brain activation patterns during a manual stimulus-response compatibility task in 18 schizophrenic patients and 18 controls. We found that across groups incongruent vs. congruent responses recruited a bilateral network consisting of dorsal fronto-parietal circuits as well as bilateral anterior insula, dorsolateral prefrontal cortex (DLPFC) and the presupplementary motor area (preSMA). When testing for the main-effect across all conditions, patients showed significantly lower activation of the right DLPFC and, in turn, increased activation in a left hemispheric network including parietal and premotor areas as well as the preSMA. For incongruent responses patients showed significantly increased activation in a similar left hemispheric network, as well as additional activation in parietal and premotor regions in the right hemisphere. The present study reveals that hypoactivity in the right DLPFC in schizophrenic patients is accompanied by hyperactivity in several fronto-parietal regions associated with task execution. Impaired top-down control due to a dysfunctional DLPFC might thus be partly compensated by an up-regulation of task-relevant regions in schizophrenic patients.

Interaction between effects of genes coding for dopamine and glutamate transmission on striatal and parahippocampal function

The genes for the dopamine transporter (DAT) and the D-Amino acid oxidase activator (DAOA or G72) have been independently implicated in the risk for schizophrenia and in bipolar disorder and/or their related intermediate phenotypes. DAT and G72 respectively modulate central dopamine and glutamate transmission, the two systems most robustly implicated in these disorders. Contemporary studies have demonstrated that elevated dopamine function is associated with glutamatergic dysfunction in psychotic disorders. Using functional magnetic resonance imaging we examined whether there was an interaction between the effects of genes that influence dopamine and glutamate transmission (DAT and G72) on regional brain activation during verbal fluency, which is known to be abnormal in psychosis, in 80 healthy volunteers. Significant interactions between the effects of G72 and DAT polymorphisms on activation were evident in the striatum, parahippocampal gyrus, and supramarginal/angular gyri bilaterally, the right insula, in the right pre-/postcentral and the left posterior cingulate/retrosplenial gyri (P < 0.05, FDR-corrected across the whole brain). This provides evidence that interactions between the dopamine and the glutamate system, thought to be altered in psychosis, have an impact in executive processing which can be modulated by common genetic variation.

Diminished activation of motor working-memory networks in Parkinson's disease

Parkinson's disease (PD) is characterized by typical extrapyramidal motor features and increasingly recognized non-motor symptoms such as working memory (WM) deficits. Using functional magnetic resonance imaging (fMRI), we investigated differences in neuronal activation during a motor WM task in 23 non-demented PD patients and 23 age- and gender-matched healthy controls. Participants had to memorize and retype variably long visuo-spatial stimulus sequences after short or long delays (immediate or delayed serial recall). PD patients showed deficient WM performance compared to controls, which was accompanied by reduced encoding-related activation in WM-related regions. Mirroring slower motor initiation and execution, reduced activation in motor structures such as the basal ganglia and superior parietal cortex was detected for both immediate and delayed recall. Increased activation in limbic, parietal and cerebellar regions was found during delayed recall only. Increased load-related activation for delayed recall was found in the posterior midline and the cerebellum. Overall, our results demonstrate that impairment of WM in PD is primarily associated with a widespread reduction of task-relevant activation, whereas additional parietal, limbic and cerebellar regions become more activated relative to matched controls. While the reduced WM-related activity mirrors the deficient WM performance, the additional recruitment may point to either dysfunctional compensatory strategies or detrimental crosstalk from “default-mode” regions, contributing to the observed impairment.

Activation shift in elderly subjects across functional systems: an fMRI study

The functional specificity of brain areas is diminished with age and accompanied by the recruitment of additional brain regions in healthy older adults. This process has repeatedly been demonstrated within distinct functional domains, in particular the visual system. However, it is yet unclear, whether this phenomenon in healthy aging, i.e., a reduced activation of task-associated areas and increased activation of additional regions, is also present across different functional systems. In the present functional imaging study, comprising 102 healthy subjects, we therefore assessed two distinct tasks engaging the sensory-motor system and the visual attention system, respectively. We found a significant interaction between age and task in the parietal operculum bilaterally. This area as a part of the sensory-motor system showed an age-related decrease in its BOLD-response to the motor task and an age-related increase of neural activity in response to the visual attention task. The opposite response pattern, i.e., reduced visual attention activation and increased response to the motor task, was observed for regions associated with the visual task: the superior parietal area 7A and the dorsal pre-motor cortex. Importantly, task performance was not correlated with age in either task. This age-by-task interaction indicates that a reduction of functional specificity in the aging brain may be counteracted by the increased recruitment of additional regions not only within, but also across functional domains. Our results thus emphasize the need for comparisons across different functional domains to gain a better understanding of age-related effects on the specificity of functional systems.

Effects of prior information on decoding degraded speech: an fMRI study

Expectations and prior knowledge are thought to support the perceptual analysis of incoming sensory stimuli, as proposed by the predictive‐coding framework. The current fMRI study investigated the effect of prior information on brain activity during the decoding of degraded speech stimuli. When prior information enabled the comprehension of the degraded sentences, the left middle temporal gyrus and the left angular gyrus were activated, highlighting a role of these areas in meaning extraction. In contrast, the activation of the left inferior frontal gyrus (area 44/45) appeared to reflect the search for meaningful information in degraded speech material that could not be decoded because of mismatches with the prior information. Our results show that degraded sentences evoke instantaneously different percepts and activation patterns depending on the type of prior information, in line with prediction‐based accounts of perception. Hum Brain Mapp 35:61–74, 2014. © 2012 Wiley Periodicals, Inc.

FMRI group analysis combining effect estimates and their variances

Conventional functional magnetic resonance imaging (FMRI) group analysis makes two key assumptions that are not always justified. First, the data from each subject is condensed into a single number per voxel, under the assumption that within-subject variance for the effect of interest is the same across all subjects or is negligible relative to the cross-subject variance. Second, it is assumed that all data values are drawn from the same Gaussian distribution with no outliers. We propose an approach that does not make such strong assumptions, and present a computationally efficient frequentist approach to FMRI group analysis, which we term mixed-effects multilevel analysis (MEMA), that incorporates both the variability across subjects and the precision estimate of each effect of interest from individual subject analyses. On average, the more accurate tests result in higher statistical power, especially when conventional variance assumptions do not hold, or in the presence of outliers. In addition, various heterogeneity measures are available with MEMA that may assist the investigator in further improving the modeling. Our method allows group effect t-tests and comparisons among conditions and among groups. In addition, it has the capability to incorporate subject-specific covariates such as age, IQ, or behavioral data. Simulations were performed to illustrate power comparisons and the capability of controlling type I errors among various significance testing methods, and the results indicated that the testing statistic we adopted struck a good balance between power gain and type I error control. Our approach is instantiated in an open-source, freely distributed program that may be used on any dataset stored in the universal neuroimaging file transfer (NIfTI) format. To date, the main impediment for more accurate testing that incorporates both within- and cross-subject variability has been the high computational cost. Our efficient implementation makes this approach practical. We recommend its use in lieu of the less accurate approach in the conventional group analysis.

Dynamic interactions in the fronto-parietal network during a manual stimulus-response compatibility task

Attentional orienting can be modulated by stimulus-driven bottom-up as well as task-dependent top-down processes. In a recent study we investigated the interaction of both processes in a manual stimulus-response compatibility task. Whereas the intraparietal sulcus (IPS) and the dorsal premotor cortex (dPMC) were involved in orienting towards the stimulus side facilitating congruent motor responses, the right temporoparietal junction (TPJ), right dorsolateral prefrontal cortex (DLPFC) as well as the preSMA sustained top-down control processes involved in voluntary reorienting. Here we used dynamic causal modelling to investigate the contributions and task-dependent interactions between these regions. Thirty-six models were tested, all of which included bilateral IPS, dPMC and primary motor cortex (M1) as a network transforming visual input into motor output as well as the right TPJ, right DLPFC and the preSMA as task-dependent top-down regions influencing the coupling within the dorsal network. Our data showed the right temporoparietal junction to play a mediating role during attentional reorienting processes by modulating the inter-hemispheric balance between both IPS. Analysis of connection strength supported the proposed role of the preSMA in controlling motor responses promoting or suppressing activity in primary motor cortex. As the results did not show a clear tendency towards a role of the right DLPFC, we propose this region, against the usual interpretation of an inhibitory influence in stimulus-response compatibility tasks, to subserve generic monitoring processes. Our DCM study hence provides evidence for context-dependent top-down control of right TPJ and DLPFC as well as the preSMA in stimulus-response compatibility.

Incongruence effects in crossmodal emotional integration

Emotions are often encountered in a multimodal fashion. Consequently, contextual framing by other modalities can alter the way that an emotional facial expression is perceived and lead to emotional conflict. Whole brain fMRI data was collected when 35 healthy subjects judged emotional expressions in faces while concurrently being exposed to emotional (scream, laughter) or neutral (yawning) sounds. The behavioral results showed that subjects rated fearful and neutral faces as being more fearful when accompanied by screams than compared to yawns (and laughs for fearful faces). Moreover, the imaging data revealed that incongruence of emotional valence between faces and sounds led to increased activation in the middle cingulate cortex, right superior frontal cortex, right supplementary motor area as well as the right temporoparietal junction. Against expectations no incongruence effects could be found in the amygdala. Further analyses revealed that, independent of emotional valence congruency, the left amygdala was consistently activated when the information from both modalities was emotional. If a neutral stimulus was present in one modality and emotional in the other, activation in the left amygdala was significantly attenuated. These results indicate that incongruence of emotional valence in audiovisual integration activates a cingulate-fronto-parietal network involved in conflict monitoring and resolution. Furthermore in audiovisual pairing amygdala responses seem to signal also the absence of any neutral feature rather than only the presence of an emotionally charged one.

Dissociating bottom-up and top-down processes in a manual stimulus-response compatibility task

Speed and accuracy of motor responses to lateralized stimuli are influenced by the spatial overlap between stimulus location and required response. Responses showing high spatial overlap with peripheral cues benefit from a bottom-up driven enhancement of attention to the respective location, whereas low overlap requires top-down modulated reorienting of resources. Here we investigated the interaction between these two processes using a spatial stimulus-response compatibility task. Subjects had to react to lateralized visual stimuli with a button press using either the ipsilateral (congruent condition) or the contralateral (incongruent condition) index finger. Stimulus-driven bottom-up processes were associated with significant contralateral activation in V5, the intraparietal sulcus (IPS) and the premotor cortex (PMC). Incongruent versus congruent responses evoked significant activation in bilateral IPS and PMC, highly overlapping with the activations found for stimulus-driven bottom-up processes, as well as additional activation in bilateral anterior insula and right dorsolateral prefrontal cortex (DLPFC) and temporoparietal junction (TPJ). Moreover, a region anterior to the bottom-up driven activation in the IPS was associated with top-down modulated directionality-specific reorienting of motor attention during incongruent motor responses. Based on these results, we propose that stimulus-driven activation of contralateral IPS and PMC represent key neuronal substrates for the behavioral advantage observed when reacting toward a congruently lateralized stimulus. Additional activation in bilateral insula and right DLPFC and TPJ during incongruent responses should reflect top-down control mechanisms mediating contextual (i.e., task) demands. Furthermore, this study provides evidence for both overlapping and disparate substrates of bottom-up and top-down modulated attentional processes in the IPS.

Simple group fMRI modeling and inference

While many advanced mixed-effects models have been proposed and are used in fMRI, the simplest, ordinary least squares (OLS), is still the one that is most widely used. A survey of 90 papers found that 92% of group fMRI analyses used OLS. Despite the widespread use, this simple approach has never been thoroughly justified and evaluated; for example, the typical reference for the method is a conference abstract, (Holmes, A., Friston, K., 1998. Generalisability, random effects & population inference. NeuroImage 7 (4 (2/3)), S754, proceedings of Fourth International Conference on Functional Mapping of the Human Brain, June 7-12, 1998, Montreal, Canada.), which has been referenced over 400 times. In this work we fully derive the simplified method in a general setting and carefully identify the homogeneity assumptions it is based on. We examine the specificity (Type I error rate) of the OLS method under heterogeneity in the one-sample case and find that the OLS method is valid, with only slight conservativeness. Surprisingly, a Satterthwaite approximation for effective degrees of freedom only makes the method more conservative, instead of more accurate. While other authors have highlighted the inferior power of the OLS method relative to optimal mixed-effects methods under heterogeneity, we revisit these results and find the power differences very modest. While statistical methods that make the best use of the data are always to be preferred, software or other practical concerns may require the use of the simple OLS group modeling. In such cases, we find that group mean inferences will be valid under the null hypothesis and will have nearly optimal sensitivity under the alternative.

Noninvasive measurement of the cerebral blood flow response in human lateral geniculate nucleus with arterial spin labeling fMRI

To date, functional magnetic resonance imaging (fMRI) studies of the lateral geniculate nucleus (LGN) have primarily focused on measures of the blood oxygenation level dependent (BOLD) signal. Arterial spin labeling (ASL) is an MRI method that can provide direct measures of functional cerebral blood flow (CBF) changes. Because CBF is a well-defined physiological quantity that contributes to BOLD contrast, CBF measures can be used to improve the quantitative interpretation of fMRI studies. However, due in part to the low intrinsic signal-to-noise ratio of the ASL method, measures of functional CBF changes in the LGN are challenging and have not previously been reported. In this study, we demonstrate the feasibility of using ASL fMRI to measure the CBF response of the LGN to visual stimulation on a 3 T MRI system. The use of background suppression and physiological noise reduction techniques allowed reliable detection of LGN activation in all five subjects studied. The measured percent CBF response during activation ranged from 40 to 100%, assuming no interaction between the left and right LGN.

Extra-classical receptive field effects measured in striate cortex with fMRI

The aim of this study was to measure the contextual influence of globally coherent motion on visual cortical responses using functional magnetic resonance imaging. Our motivation was to test a prediction from representational theories of perception (i.e. predictive coding) that primary visual responses should be suppressed by top-down influences during coherent motion. We used a sparse stimulus array such that each element could not fall within the same classical receptive field of primary visual cortex neurons (i.e. precluding lateral interactions within V1). This enabled us to attribute differences, in striate cortex responses, to extra-classical receptive field effects mediated by backward connections. In accord with theoretical predictions we were able to demonstrate suppression of striate cortex activations to coherent relative to incoherent motion. These results suggest that suppression of primary visual cortex responses to coherent motion reflect extra-classical effects mediated by backward connections.

Two approaches to repetition suppression

Repetition suppression refers to the phenomenon that prior processing of stimuli (or stimulus attributes) decreases activation elicited by processing subsequent stimuli with identical attributes. We present two complementary approaches to identify regions that show repetition suppression for subsequent sentences with either identical: (1) sentence forms or (2) speakers. The first categorical approach simply compares sentences that are presented in Same and Different blocks. The second factorial approach operationally defines repetition suppression as decreased activation for the subsequent Same stimulus relative to its preceding sentence. To account for nonspecific time confounds, this approach tests for a repetition x condition (Same or Different) interaction. Surprisingly, the two approaches revealed different results: Only the categorical analysis detected sentence repetition effects in multiple regions within a bilateral frontotemporal system that has previously been implicated in sentence processing. These discrepancies might be due to the different efficiencies with which the particular contrasts were estimated or spurious differences in stimuli or attentional set that could not be entirely controlled within a single subject. Finally, we combined the two approaches in a [global null] conjunction analysis.

Pupillary contagion: central mechanisms engaged in sadness processing

Empathic responses underlie our ability to share emotions and sensations with others. We investigated whether observed pupil size modulates our perception of other's emotional expressions and examined the central mechanisms modulated by incidental perception of pupil size in emotional facial expressions. We show that diminishing pupil size enhances ratings of emotional intensity and valence for sad, but not happy, angry or neutral facial expressions. This effect was associated with modulation of neural activity within cortical and subcortical regions implicated in social cognition. In an identical context, we show that the observed pupil size was mirrored by the observers' own pupil size. This empathetic contagion engaged the brainstem pupillary control nuclei (Edinger-Westphal) in proportion to individual subject's sensitivity to this effect. These findings provide evidence that perception-action mechanisms extend to non-volitional operations of the autonomic nervous system.

Physiological noise reduction for arterial spin labeling functional MRI

Three methods for the reduction of physiological noise in arterial spin labeling (ASL) functional magnetic resonance imaging (fMRI) are presented and compared. The methods are based upon a general linear model of the ASL measurement process and on a previously described retrospective image-based method (RETROICOR) for physiological noise reduction in blood oxygenation level dependent fMRI. In the first method, the contribution of physiological noise to the interleaved control and tag images that comprise the ASL time series are assumed to be equal, while in the second method this assumption is not made. For the third method, it is assumed that physiological noise primarily impacts the perfusion time series obtained from the filtered subtraction of the control and tag images. The methods were evaluated using studies of functional activity in the visual cortex and the hippocampal region. The first and second methods significantly improved statistical performance in both brain regions, whereas the third method did not provide a significant gain. The second method provided significantly better performance than the first method in the hippocampal region, whereas the differences between methods were less pronounced in visual cortex. The improved performance of the second method in the hippocampal region appears to reflect the relatively greater effect of cardiac fluctuations in this brain region. The proposed methods should be particularly useful for ASL studies of cognitive processes where the intrinsic signal to noise ratio is typically lower than for studies of primary sensory regions.

A Bayesian approach to modeling dynamic effective connectivity with fMRI data

A state-space modeling approach for examining dynamic relationship between multiple brain regions was proposed in Ho, Ombao and Shumway (Ho, M.R., Ombao, H., Shumway, R., 2005. A State-Space Approach to Modelling Brain Dynamics to Appear in Statistica Sinica). Their approach assumed that the quantity representing the influence of one neuronal system over another, or effective connectivity, is time-invariant. However, more and more empirical evidence suggests that the connectivity between brain areas may be dynamic which calls for temporal modeling of effective connectivity. A Bayesian approach is proposed to solve this problem in this paper. Our approach first decomposes the observed time series into measurement error and the BOLD (blood oxygenation level-dependent) signals. To capture the complexities of the dynamic processes in the brain, region-specific activations are subsequently modeled, as a linear function of the BOLD signals history at other brain regions. The coefficients in these linear functions represent effective connectivity between the regions under consideration. They are further assumed to follow a random walk process so to characterize the dynamic nature of brain connectivity. We also consider the temporal dependence that may be present in the measurement errors. ML-II method (Berger, J.O., 1985. Statistical Decision Theory and Bayesian Analysis (2nd ed.). Springer, New York) was employed to estimate the hyperparameters in the model and Bayes factor was used to compare among competing models. Statistical inference of the effective connectivity coefficients was based on their posterior distributions and the corresponding Bayesian credible regions (Carlin, B.P., Louis, T.A., 2000. Bayes and Empirical Bayes Methods for Data Analysis (2nd ed.). Chapman and Hall, Boca Raton). The proposed method was applied to a functional magnetic resonance imaging data set and results support the theory of attentional control network and demonstrate that this network is dynamic in nature.

Mixed-effects and fMRI studies

This note concerns mixed-effect (MFX) analyses in multisession functional magnetic resonance imaging (fMRI) studies. It clarifies the relationship between mixed-effect analyses and the two-stage "summary statistics" procedure (Holmes, A.P., Friston, K.J., 1998. Generalisability, random effects and population inference. NeuroImage 7, S754) that has been adopted widely for analyses of fMRI data at the group level. We describe a simple procedure, based on restricted maximum likelihood (ReML) estimates of covariance components, that enables full mixed-effects analyses in the context of statistical parametric mapping. Using this procedure, we compare the results of a full mixed-effects analysis with those obtained from the simpler two-stage procedure and comment on the situations when the two approaches may give different results.

Optimal spatial regularisation of autocorrelation estimates in fMRI analysis

In the General Linear Model (GLM) framework for the statistical analysis of fMRI data, the problem of temporal autocorrelations in the residual signal (after regression) has been frequently addressed in the open literature. There exist various methods for correcting the ensuing bias in the statistical testing, among which the prewhitening strategy, which uses a prewhitening matrix for rendering the residual signal white (i.e., without temporal autocorrelations). This correction is only exact when the autocorrelation structure of the noise-generating process is accurately known, and the estimates derived from the fMRI data are too noisy to be used for correction. Recently, Worsley and co-workers proposed to spatially smooth the noisy autocorrelation estimates, effectively reducing their variance and allowing for a better correction. In this article, a systematic study into the effect of the smoothing kernel width is performed and a method is introduced for choosing this bandwidth in an "optimal" manner. Several aspects of the prewhitening strategy are investigated, namely the choice of the autocorrelation estimate (biased or unbiased), the accuracy of the estimates, the degree of spatial regularisation and the order of the autoregressive model used for characterising the noise. The proposed method is extensively evaluated on both synthetic and real fMRI data.

Statistical parametric mapping for event-related potentials (II): a hierarchical temporal model

In this paper, we describe a temporal model for event-related potentials (ERP) in the context of statistical parametric mapping (SPM). In brief, we project channel data onto a two-dimensional scalp surface or into three-dimensional brain space using some appropriate inverse solution. We then treat the spatiotemporal data in a mass-univariate fashion. This implicitly factorises the model into spatial and temporal components. The key contribution of this paper is the use of observation models that afford an explicit distinction between observation error and variation in the expression of ERPs. This distinction is created by employing a two-level hierarchical model, in which the first level models the ERP effects within-subject and trial type, while the second models differences in ERP expression among trial types and subjects. By bringing the analysis of ERP data into a classical hierarchical (i.e., mixed effects) framework, many apparently disparate approaches (e.g., conventional P300 analyses and time-frequency analyses of stimulus-locked oscillations) can be reconciled within the same estimation and inference procedure. Inference proceeds in the normal way using t or F statistics to test for effects that are localised in peristimulus time or in some time-frequency window. The use of F statistics is an important generalisation of classical approaches, because it allows one to test for effects that lie in a multidimensional subspace (i.e., of unknown but constrained form). We describe the analysis procedures, the underlying theory and compare its performance to established techniques.