The impact of physiological noise correction on fMRI at 7 T.

Cognitive neuroimaging studies typically require fast whole brain image acquisition with maximal sensitivity to small BOLD signal changes. To increase the sensitivity, higher field strengths are often employed, since they provide an increased image signal-to-noise ratio (SNR). However, as image SNR increases, the relative contribution of physiological noise to the total time series noise will be greater compared to that from thermal noise. At 7 T, we studied how the physiological noise contribution can be best reduced for EPI time series acquired at three different spatial resolutions (1.1 mm × 1.1 mm × 1.8 mm, 2 mm × 2 mm × 2 mm and 3 mm × 3 mm × 3 mm). Applying optimal physiological noise correction methods improved temporal SNR (tSNR) and increased the numbers of significantly activated voxels in fMRI visual activation studies for all sets of acquisition parameters. The most dramatic results were achieved for the lowest spatial resolution, an acquisition parameter combination commonly used in cognitive neuroimaging which requires high functional sensitivity and temporal resolution (i.e. 3mm isotropic resolution and whole brain image repetition time of 2s). For this data, physiological noise models based on cardio-respiratory information improved tSNR by approximately 25% in the visual cortex and 35% sub-cortically. When the time series were additionally corrected for the residual effects of head motion after retrospective realignment, the tSNR was increased by around 58% in the visual cortex and 71% sub-cortically, exceeding tSNR ~140. In conclusion, optimal physiological noise correction at 7 T increases tSNR significantly, resulting in the highest tSNR per unit time published so far. This tSNR improvement translates into a significant increase in BOLD sensitivity, facilitating the study of even subtle BOLD responses.

Estimating the transfer function from neuronal activity to BOLD using simultaneous EEG-fMRI.

Previous studies using combined electrical and hemodynamic measurements of brain activity, such as EEG and (BOLD) fMRI, have yielded discrepant results regarding the relationship between neuronal activity and the associated BOLD response. In particular, some studies suggest that this link, or transfer function, depends on the frequency content of neuronal activity, while others suggest that total neuronal power accounts for the changes in BOLD. Here we explored this dependency by comparing different frequency-dependent and -independent transfer functions, using simultaneous EEG-fMRI. Our results suggest that changes in BOLD are indeed associated with changes in the spectral profile of neuronal activity and that these changes do not arise from one specific spectral band. Instead they result from the dynamics of the various frequency components together, in particular, from the relative power between high and low frequencies. Understanding the nature of the link between neuronal activity and BOLD plays a crucial role in improving the interpretability of BOLD images as well as on the design of more robust and realistic models for the integration of EEG and fMRI.

Activity in human V1 follows multisensory perception.

When a single brief visual flash is accompanied by two auditory bleeps, it is frequently perceived incorrectly as two flashes. Such illusory multisensory perception is associated with increased activation of retinotopic human primary visual cortex (V1) suggesting that such activity reflects subjective perception [Watkins, S., Shams, L., Tanaka, S., Haynes, J.D., Rees, G., 2006. Sound alters activity in human V1 in association with illusory visual perception. Neuroimage. 31, 1247-1256]. However, an alternate possibility is that increased V1 activity reflects either fluctuating attention or auditory-visual perceptual matching on illusion trials. Here, we rule out these possibilities by studying the complementary illusion, where a double flash is accompanied by a single bleep and perceived incorrectly as a single flash. We replicate findings of increased activity in retinotopic V1 when a single flash is perceived incorrectly as two flashes, and now show that activity is decreased in retinotopic V1 when a double flash is perceived incorrectly as a single flash. Our findings provide strong support for the notion that human V1 activity reflects subjective perception in these multisensory illusions.

Action selectivity in parietal and temporal cortex.

The sensory-action theory proposes that the neural substrates underlying action representations are related to a visuomotor action system encompassing the left ventral premotor cortex, the anterior intraparietal (AIP) and left posterior middle temporal gyrus (LPMT). Using fMRI, we demonstrate that semantic decisions on action, relative to non-action words, increased activation in the left AIP and LPMT irrespective of whether the words were presented in a written or spoken form. Left AIP and LPMT might thus play the role of amodal semantic regions that can be activated via auditory as well as visual input. Left AIP and LPMT did not distinguish between different types of actions such as hand actions and whole body movements, although a right STS region responded selectively to whole body movements.

Functional magnetic resonance imaging of proactive interference during spoken cued recall.

Though lesions to frontal cortex can increase susceptibility to interference from previously established but irrelevant memories ("proactive interference"), the specific regions underlying this problem are difficult to determine because the lesions are typically large and heterogeneous. We used event-related functional magnetic resonance imaging to investigate proactive interference in healthy volunteers performing an "AB-AC" paired-associate cued-recall paradigm. At Study, participants intentionally encoded semantically related visual word pairs, which were changed three times (high interference), repeated three times (low interference), or presented only once. At Test, participants were presented with the first word of each pair and attempted to recall its most recent associate from the Study phase. To overcome the problem of image artifacts caused by speech-related head motion, we cued speech during a gap between image acquisitions. Regions in left inferior frontal cortex and bilateral frontopolar cortex showed interference effects during both Study and Test. The pattern of responses in these regions differed, however. Left inferior frontal regions showed mainly reduced responses associated with low interference, whereas frontopolar regions showed mainly increased responses associated with high interference. When incorrect as well as correct trials were analyzed at Test, additional activation associated with high interference was observed in right dorsolateral prefrontal cortex. These data suggest that distinct regions within prefrontal cortex subserve different functions in the presence of proactive interference during cued recall.

Compensation of susceptibility-induced BOLD sensitivity losses in echo-planar fMRI imaging.

Gradient-echo echo-planar imaging is a standard technique in functional magnetic resonance imaging (fMRI) experiments based on the blood oxygenation level-dependent (BOLD) effect. A major problem is the occurrence of susceptibility gradients near air/tissue interfaces. As a consequence, the detection of neuronal activation may be greatly compromised in certain brain areas, especially in the temporal lobes and in the orbitofrontal cortex. Common approaches to overcome this problem, such as z-shimming or the use of tailored radio frequency pulses, usually compensate only for susceptibility gradients in the slice selection direction. In the present study, the influence of susceptibility gradients in the phase encoding direction is investigated both theoretically and experimentally. It is shown that these gradients influence the effective echo time TE and may reduce considerably the local BOLD sensitivity, even in the case of acceptable image intensities. A compensation method is proposed and tested in an fMRI experiment based on a hypercapnic challenge. The results suggest that the compensation method allows for the detection of activation in brain areas which are usually unavailable for BOLD studies.

Dissociable human perirhinal, hippocampal, and parahippocampal roles during verbal encoding.

The precise contribution of perirhinal cortex to human episodic memory is uncertain. Human intracranial recordings highlight a role in successful episodic memory encoding, but encoding-related perirhinal activation has not been observed with functional imaging. By adapting functional magnetic resonance imaging scanning parameters to maximize sensitivity to medial temporal lobe activity, we demonstrate that left perirhinal and hippocampal responses during word list encoding are greater for subsequently recalled than forgotten words. Although perirhinal responses predict memory for all words, successful encoding of initial words in a list, demonstrating a primacy effect, is associated with parahippocampal and anterior hippocampal activation. We conclude that perirhinal cortex and hippocampus participate in successful memory encoding. Encoding-related parahippocampal and anterior hippocampal responses for initial, remembered words most likely reflects enhanced attentional orienting to these positionally distinctive items.

Learning arbitrary visuomotor associations: temporal dynamic of brain activity.

Primates can give behavioral responses on the basis of arbitrary, context-dependent rules. When sensory instructions and behavioral responses are associated by arbitrary rules, these rules need to be learned. This study investigates the temporal dynamics of functional segregation at the basis of visuomotor associative learning in humans, isolating specific learning-related changes in neurovascular activity across the whole brain. We have used fMRI to measure human brain activity during performance of two tasks requiring the association of visual patterns with motor responses. Both tasks were learned by trial and error, either before (visuomotor control) or during (visuomotor learning) the scanning session. Epochs of tasks performance ( approximately 30 s) were alternated with a baseline period over the whole scanning session ( approximately 50 min). We have assessed both linear and nonlinear modulations in the differential signal between tasks, independently from overall task differences. The performance indices of the visuomotor learning task smoothly converged onto the values of a steady-state control condition, according to nonlinear timecourses. Specific visuomotor learning-related activity has been found over a distributed cortical network, centred on a temporo-prefrontal circuit. These cortical time-modulated activities were supported early in learning by the hippocampal/parahippocampal complex, and late in learning by the basal ganglia system. These findings suggest the inferior temporal and the ventral prefrontal cortex are critical neural nodes for integrating perceptual information with executive processes.

Event-related fMRI with simultaneous and continuous EEG: description of the method and initial case report.

We report on the initial imaging findings with a new technique for the simultaneous and continuous acquisition of functional MRI data and EEG recording. Thirty-seven stereotyped interictal epileptiform discharges (spikes) were identified on EEG recorded continuously during the fMRI acquisition on a patient with epilepsy. Localization of the BOLD activation associated with the EEG events was consistent with previous findings and EEG source modeling. The time course of activation was comparable with the physiological hemodynamic response function (HRF). The new methodology could lead to novel and important applications in many areas of neuroscience.

Does hypercapnia-induced cerebral vasodilation modulate the hemodynamic response to neural activation?

Increases in cerebral blood flow produced by vasoactive agents will increase blood oxygen level-dependent (BOLD) MRI signal intensity. The effects of such vasodilation on activation-related signal changes are incompletely characterized. The two signal changes may be simply additive or there may be more a complex interaction. To investigate this, BOLD MRI was performed in four normal male subjects using T2*-weighted echo planar imaging; brain volumes were acquired every 6.2 s, using a Siemens VISION scanner operating at 2 Tesla; each volume consisted of 64 sequential transverse slices (64 x 64 pixels per slice, 3 x 3 x 3 mm). Sixteen periods of visual stimulation were produced using a flickering checkerboard (8 Hz, 31 s On/31 s Off); this was coupled with five periods of hypercapnia (4% inspired CO(2), 62 s On/124 s Off). Data were analyzed using SPM96. Mean signal intensity, calculated globally for the whole brain, closely mirrored changes in the partial pressure of end-tidal CO(2) (PCO(2)), and hypercapnia was associated with widespread significant signal increases (P < 0.001), predominantly within grey matter. As expected, the visual stimulation produced significant signal changes within the occipital cortex (P < 0.001). Within the occipital cortex, no significant interactions (P > 0.001) between the effects of the visual stimulation and PCO(2) were present. The increases in PCO(2) imposed dynamically in the present study would increase cerebral blood flow by between 25 and 40%, an increase within the physiological range and comparable to that induced by neural activation. With this flow change the effects of vasodilation, on an activation-related signal change, are simply additive.

Encoding of the temporal regularity of sound in the human brainstem.

We measured the neural activity associated with the temporal structure of sound in the human auditory pathway from cochlear nucleus to cortex. The temporal structure includes regularities at the millisecond level and pitch sequences at the hundreds-of-milliseconds level. Functional magnetic resonance imaging (fMRI) of the whole brain with cardiac triggering allowed simultaneous observation of activity in the brainstem, thalamus and cerebrum. This work shows that the process of recoding temporal patterns into a more stable form begins as early as the cochlear nucleus and continues up to auditory cortex.

Learning- and expectation-related changes in the human brain during motor learning.

We have studied a simple form of motor learning in the human brain so as to isolate activity related to motor learning and the prediction of sensory events. Whole-brain, event-related functional magnetic resonance imaging (fMRI) was used to record activity during classical discriminative delay eyeblink conditioning. Auditory conditioned stimulus (CS+) trials were presented either with a corneal airpuff unconditioned stimulus (US, paired), or without a US (unpaired). Auditory CS- trials were never reinforced with a US. Trials were presented pseudorandomly, 66 times each. The subjects gradually produced conditioned responses to CS+ trials, while increasingly differentiating between CS+ and CS- trials. The increasing difference between hemodynamic responses for unpaired CS+ and for CS- trials evolved slowly during conditioning in the ipsilateral cerebellar cortex (Crus I/Lobule HVI), contralateral motor cortex and hippocampus. To localize changes that were related to sensory prediction, we compared trials on which the expected airpuff US failed to occur (Unpaired CS+) with trials on which it occurred as expected (Paired CS+). Error-related signals in the contralateral cerebellum and somatosensory cortex were seen to increase during learning as the sensory prediction became stronger. The changes seen in the ipsilateral cerebellar cortex may be due either to the violations of sensory predictions, or to learning-related increases in the excitability of cerebellar neurons to presentations of the CS+.

Optimization of 3-D MP-RAGE sequences for structural brain imaging.

An optimized MR sequence for structural three-dimensional brain scans is presented, giving good T(1) contrast and excellent white matter/gray matter segmentation. Modification of the usual linear phase encoding order to centric phase encoding restores the contrast loss, which usually occurs after magnetization preparation during the acquisition process when large volumes are imaged. The deleterious effects on the point-spread function are compensated by means of an appropriate k-space filter. RF coil inhomogeneities are corrected by means of shaped excitation pulses. High contrast-to-noise images of the entire brain with 1 mm isotropic resolution can be obtained in 12 min. The contrast-to-noise-ratio is about 100% higher than for sequences based on linear phase encoding.

To smooth or not to smooth? Bias and efficiency in fMRI time-series analysis.

This paper concerns temporal filtering in fMRI time-series analysis. Whitening serially correlated data is the most efficient approach to parameter estimation. However, if there is a discrepancy between the assumed and the actual correlations, whitening can render the analysis exquisitely sensitive to bias when estimating the standard error of the ensuing parameter estimates. This bias, although not expressed in terms of the estimated responses, has profound effects on any statistic used for inference. The special constraints of fMRI analysis ensure that there will always be a misspecification of the assumed serial correlations. One resolution of this problem is to filter the data to minimize bias, while maintaining a reasonable degree of efficiency. In this paper we present expressions for efficiency (of parameter estimation) and bias (in estimating standard error) in terms of assumed and actual correlation structures in the context of the general linear model. We show that: (i) Whitening strategies can result in profound bias and are therefore probably precluded in parametric fMRI data analyses. (ii) Band-pass filtering, and implicitly smoothing, has an important role in protecting against inferential bias.

A method for removing imaging artifact from continuous EEG recorded during functional MRI.

Combined EEG/fMRI recording has been used to localize the generators of EEG events and to identify subject state in cognitive studies and is of increasing interest. However, the large EEG artifacts induced during fMRI have precluded simultaneous EEG and fMRI recording, restricting study design. Removing this artifact is difficult, as it normally exceeds EEG significantly and contains components in the EEG frequency range. We have developed a recording system and an artifact reduction method that reduce this artifact effectively. The recording system has large dynamic range to capture both low-amplitude EEG and large imaging artifact without distortion (resolution 2 microV, range 33.3 mV), 5-kHz sampling, and low-pass filtering prior to the main gain stage. Imaging artifact is reduced by subtracting an averaged artifact waveform, followed by adaptive noise cancellation to reduce any residual artifact. This method was validated in recordings from five subjects using periodic and continuous fMRI sequences. Spectral analysis revealed differences of only 10 to 18% between EEG recorded in the scanner without fMRI and the corrected EEG. Ninety-nine percent of spike waves (median 74 microV) added to the recordings were identified in the corrected EEG compared to 12% in the uncorrected EEG. The median noise after artifact reduction was 8 microV. All these measures indicate that most of the artifact was removed, with minimal EEG distortion. Using this recording system and artifact reduction method, we have demonstrated that simultaneous EEG/fMRI studies are for the first time possible, extending the scope of EEG/fMRI studies considerably.

The prefrontal cortex: response selection or maintenance within working memory?

It is controversial whether the dorsolateral prefrontal cortex is involved in the maintenance of items in working memory or in the selection of responses. We used event-related functional magnetic resonance imaging to study the performance of a spatial working memory task by humans. We distinguished the maintenance of spatial items from the selection of an item from memory to guide a response. Selection, but not maintenance, was associated with activation of prefrontal area 46 of the dorsal lateral prefrontal cortex. In contrast, maintenance was associated with activation of prefrontal area 8 and the intraparietal cortex. The results support a role for the dorsal prefrontal cortex in the selection of representations. This accounts for the fact that this area is activated both when subjects select between items on working memory tasks and when they freely select between movements on tasks of willed action.

EEG recording during fMRI experiments: image quality.

Electroencephalographic (EEG) monitoring during functional magnetic resonance imaging (fMRI) experiments is increasingly applied for studying physiological and pathological brain function. However, the quality of the fMRI data can be significantly compromised by the EEG recording due to the magnetic susceptibility of the EEG electrode assemblies and electromagnetic noise emitted by the EEG recording equipment. We therefore investigated the effect of individual components of the EEG recording equipment on the quality of echo planar images. The artifact associated with each component was measured and compared to the minimum scalp-cortex distance measured in normal controls. The image noise originating from the EEG recording equipment was identified as coherent noise and could be eliminated by appropriate shielding of the EEG equipment. It was concluded that concurrent EEG and fMRI could be performed without compromising the image quality significantly if suitable equipment is used. The methods described and the results of this study should be useful to other researchers as a framework for testing of their own equipment and for the selection of appropriate equipment for EEG recording inside a MR scanner.

Characterization and correction of interpolation effects in the realignment of fMRI time series.

Subject motion in functional magnetic resonance imaging (fMRI) studies can be accurately estimated using realignment algorithms. However, residual changes in signal intensity arising from motion have been identified in the data even after realignment of the image time series. The nature of these artifacts is characterized using simulated displacements of an fMRI image and is attributed to interpolation errors introduced by the resampling inherent within realignment. A correction scheme that uses a periodic function of the estimated displacements to remove interpolation errors from the image time series on a voxel-by-voxel basis is proposed. The artifacts are investigated using a brain phantom to avoid physiological confounds. Small- and large-scale systematic displacements show that the artifacts have the same form as revealed by the simulated displacements. A randomly displaced phantom and a human subject are used to demonstrate that interpolation errors are minimized using the correction.

Auditory processing across the sleep-wake cycle: simultaneous EEG and fMRI monitoring in humans.

We combined fMRI and EEG recording to study the neurophysiological responses associated with auditory stimulation across the sleep-wake cycle. We found that presentation of auditory stimuli produces bilateral activation in auditory cortex, thalamus, and caudate during both wakefulness and nonrapid eye movement (NREM) sleep. However, the left parietal and, bilaterally, the prefrontal and cingulate cortices and the thalamus were less activated during NREM sleep compared to wakefulness. These areas may play a role in the further processing of sensory information required to achieve conscious perception during wakefulness. Finally, during NREM sleep, the left amygdala and the left prefrontal cortex were more activated by stimuli having special affective significance than by neutral stimuli. These data suggests that the sleeping brain can process auditory stimuli and detect meaningful events.

Stochastic designs in event-related fMRI.

This article considers the efficiency of event-related fMRI designs in terms of the optimum temporal pattern of stimulus or trial presentations. The distinction between "stochastic" and "deterministic" is used to distinguish between designs that are specified in terms of the probability that an event will occur at a series of time points (stochastic) and those in which events always occur at prespecified time (deterministic). Stochastic designs may be "stationary," in which the probability is constant, or nonstationary, in which the probabilities change with time. All these designs can be parameterized in terms of a vector of occurrence probabilities and a prototypic design matrix that embodies constraints (such as the minimum stimulus onset asynchrony) and the model of hemodynamic responses. A simple function of these parameters is presented and used to compare the relative efficiency of different designs. Designs with slow modulation of occurrence probabilities are generally more efficient than stationary designs. Interestingly the most efficient design is a conventional block design. A critical point, made in this article, is that the most efficient design for one effect may not be the most efficient for another. This is particularly important when considering evoked responses and the differences among responses. The most efficient designs for evoked responses, as opposed to differential responses, require trial-free periods during which baseline levels can be attained. In the context of stochastic, rapid-presentation designs this is equivalent to the inclusion of "null events."