Adaptive changes in early and late blind: a fMRI study of Braille reading.

Braille reading depends on remarkable adaptations that connect the somatosensory system to language. We hypothesized that the pattern of cortical activations in blind individuals reading Braille would reflect these adaptations. Activations in visual (occipital-temporal), frontal-language, and somatosensory cortex in blind individuals reading Braille were examined for evidence of differences relative to previously reported studies of sighted subjects reading print or receiving tactile stimulation. Nine congenitally blind and seven late-onset blind subjects were studied with fMRI as they covertly performed verb generation in response to reading Braille embossed nouns. The control task was reading the nonlexical Braille string "######". This study emphasized image analysis in individual subjects rather than pooled data. Group differences were examined by comparing magnitudes and spatial extent of activated regions first determined to be significant using the general linear model. The major adaptive change was robust activation of visual cortex despite the complete absence of vision in all subjects. This included foci in peri-calcarine, lingual, cuneus and fusiform cortex, and in the lateral and superior occipital gyri encompassing primary (V1), secondary (V2), and higher tier (VP, V4v, LO and possibly V3A) visual areas previously identified in sighted subjects. Subjects who never had vision differed from late blind subjects in showing even greater activity in occipital-temporal cortex, provisionally corresponding to V5/MT and V8. In addition, the early blind had stronger activation of occipital cortex located contralateral to the hand used for reading Braille. Responses in frontal and parietal cortex were nearly identical in both subject groups. There was no evidence of modifications in frontal cortex language areas (inferior frontal gyrus and dorsolateral prefrontal cortex). Surprisingly, there was also no evidence of an adaptive expansion of the somatosensory or primary motor cortex dedicated to the Braille reading finger(s). Lack of evidence for an expected enlargement of the somatosensory representation may have resulted from balanced tactile stimulation and gross motor demands during Braille reading of nouns and the control fields. Extensive engagement of visual cortex without vision is discussed in reference to the special demands of Braille reading. It is argued that these responses may represent critical language processing mechanisms normally present in visual cortex.

Increased amygdala response to masked emotional faces in depressed subjects resolves with antidepressant treatment: an fMRI study.

BACKGROUND: The amygdala has a central role in processing emotions, particularly fear. During functional magnetic resonance imaging (fMRI) amygdala activation has been demonstrated outside of conscious awareness using masked emotional faces. METHODS: We applied the masked faces paradigm to patients with major depression (n = 11) and matched control subjects (n = 11) during fMRI to compare amygdala activation in response to masked emotional faces before and after antidepressant treatment. Data were analyzed using left and right amygdala a priori regions of interest, in an analysis of variance block analysis and random effects model. RESULTS: Depressed patients had exaggerated left amygdala activation to all faces, greater for fearful faces. Right amygdala did not differ from control subjects. Following treatment, patients had bilateral reduced amygdala activation to masked fearful faces and bilateral reduced amygdala activation to all faces. Control subjects had no differences between the two scanning sessions. CONCLUSIONS: Depressed patients have left amygdala hyperarousal, even when processing stimuli outside conscious awareness. Increased amygdala activation normalizes with antidepressant treatment.

Direct comparison of prefrontal cortex regions engaged by working and long-term memory tasks.

Neuroimaging studies have suggested the involvement of ventrolateral, dorsolateral, and frontopolar prefrontal cortex (PFC) regions in both working (WM) and long-term memory (LTM). The current study used functional magnetic resonance imaging (fMRI) to directly compare whether these PFC regions show selective activation associated with one memory domain. In a within-subjects design, subjects performed the n-back WM task (two-back condition) as well as LTM encoding (intentional memorization) and retrieval (yes-no recognition) tasks. Additionally, each task was performed with two different types of stimulus materials (familiar words, unfamiliar faces) in order to determine the influence of material-type vs task-type. A bilateral region of dorsolateral PFC (DL-PFC; BA 46/9) was found to be selectively activated during the two-back condition, consistent with a hypothesized role for this region in active maintenance and/or manipulation of information in WM. Left frontopolar PFC (FP-PFC) was also found to be selectively engaged during the two-back. Although FP-PFC activity has been previously associated with retrieval from LTM, no frontopolar regions were found to be selectively engaged by retrieval. Finally, lateralized ventrolateral PFC (VL-PFC) regions were found to be selectively engaged by material-type, but uninfluenced by task-type. These results highlight the importance of examining PFC activity across multiple memory domains, both for functionally differentiating PFC regions (e.g., task-selectivity vs material-selectivity in DL-PFC and VL-PFC) and for testing the applicability of memory domain-specific theories (e.g., FP-PFC in LTM retrieval).

Human brain activity time-locked to perceptual event boundaries.

Temporal structure has a major role in human understanding of everyday events. Observers are able to segment ongoing activity into temporal parts and sub-parts that are reliable, meaningful and correlated with ecologically relevant features of the action. Here we present evidence that a network of brain regions is tuned to perceptually salient event boundaries, both during intentional event segmentation and during naive passive viewing of events. Activity within this network may provide a basis for parsing the temporally evolving environment into meaningful units.

Dissociating state and item components of recognition memory using fMRI.

Cognitive functions such as memory retrieval involve a combination of state- and item-related processes. State-related processes are sustained throughout a task (e.g., "retrieval mode" associated with ongoing goals), whereas item-related processes are transient and allied to individual stimuli (e.g., "retrieval success" associated with the recovery of information from memory). The present study employed a mixed "blocked and event-related" experimental design to identify neural mechanisms that support state- and item-related processes during a recognition memory task. Subjects alternated between blocks of fixation and recognition memory (discriminating between old and new words). Critically, event-related procedures were embedded within the recognition blocks, including the jittering of sequential trials. This design ensures that the temporal profiles of state- and item-related activity differ and consequently renders them separable; without this procedure item-related activity would summate to produce a state-like response. Results suggest three classes of brain region support recognition memory, exhibiting: (1) predominantly transient activity (including regions in medial parietal, lateral parietal, and anterior left frontal cortex) reflecting item-related processing associated with "retrieval success," (2) predominantly sustained activity (including decreased activity in bilateral parahippocampal cortex) reflecting state-related processing associated with "retrieval mode," (3) concurrent sustained and transient activity (including regions in left middle frontal gyrus, bilateral frontal operculum, and medial frontal gyrus), reflecting a combination of state- and item-related processing. The present findings support the idea that recognition memory tasks are dependent upon a combination of state- and item-related processes that have dissociable neural correlates identifiable using fMRI. Moreover, the mixed "blocked and event-related" design employed here provides a general procedure for separating state- and item-related processes.

Separating processes within a trial in event-related functional MRI I. The Method.

Many behavioral paradigms involve temporally overlapping sensory, cognitive, and motor components within a single trial. The complex interplay among these factors makes it desirable to separate the components of the total response without assumptions about shape of the underlying hemodynamic response. We present a method that does this. Four conditions were studied in four subjects to validate the method. Two conditions involved rapid event-related studies, one with a low-contrast (5%) flickering checkerboard and another with a high-contrast (95%) checkerboard. In the third condition, the same high-contrast checkerboard was presented with widely spaced trials. Finally, multicomponent trials were formed from temporally adjacent low-contrast and high-contrast stimuli. These trials were presented as a rapid event-related study. Low-contrast stimuli presented in isolation (partial trials) made it possible to uniquely estimate both the low-contrast and high-contrast responses. These estimated responses matched those measured in the first three conditions, thereby validating the method. Nonlinear interactions between adjacent low-contrast and high-contrast responses were shown to be significant but weak in two of the four subjects.

Separating processes within a trial in event-related functional MRI II. Analysis.

Many cognitive processes occur on time scales that can significantly affect the shape of the blood oxygenation level-dependent (BOLD) response in event-related functional MRI. This shape can be estimated from event related designs, even if these processes occur in a fixed temporal sequence (J. M. Ollinger, G. L. Shulman, and M. Corbetta. 2001. NeuroImage 13: 210-217). Several important considerations come into play when interpreting these time courses. First, in single subjects, correlations among neighboring time points give the noise a smooth appearance that can be confused with changes in the BOLD response. Second, the variance and degree of correlation among estimated time courses are strongly influenced by the timing of the experimental design. Simulations show that optimal results are obtained if the intertrial intervals are as short as possible, if they follow an exponential distribution with at least three distinct values, and if 40% of the trials are partial trials. These results are not particularly sensitive to the fraction of partial trials, so accurate estimation of time courses can be obtained with lower percentages of partial trials (20-25%). Third, statistical maps can be formed from F statistics computed with the extra sum of square principle or by t statistics computed from the cross-correlation of the time courses with a model for the hemodynamic response. The latter method relies on an accurate model for the hemodynamic response. The most robust model among those tested was a single gamma function. Finally, the power spectrum of the measured BOLD signal in rapid event-related paradigms is similar to that of the noise. Nevertheless, high-pass filtering is desirable if the appropriate model for the hemodynamic response is used.

Multiple neural correlates of detection in the human brain.

We used event-related functional MRI to examine the neural consequences of detecting the presence or absence of a stimulus. Subjects detected a brief interval of coherent motion embedded in dynamic noise that was presented throughout a test period. Several brain regions, including V1/V2, middle temporal complex (MT+), left intraparietal cortex, and the frontal eye field, were activated at the onset of the dynamic noise, irrespective of whether a coherent motion target was presented early or late in the test period, or not at all. These regions, many of which were motion sensitive, were likely involved in searching for and detecting the target. The blood oxygenation level-dependent signal in these regions was higher in trials in which a target was detected than in trials in which it was missed or not presented, indicating that these regions were modulated by detection. Moreover, the blood oxygenation leveldependent signal in these regions decayed quickly once a target was detected, even though the dynamic noise continued to be displayed, indicating that they were shut down after detection. Therefore, detection-related modulations occurred in the same regions that accumulate target information over time, in agreement with current psychological and neural models of detection. Many other regions, however, including areas in prefrontal cortex and anterior cingulate, were not involved in searching for a target. In these regions, activation began early in the test period when an early target was detected but began late in the test period when a late target was detected or when a response was correctly withheld in the absence of a motion target. The signal in these regions was therefore triggered by a discrete event during the test interval that was related to presence-absence detection.

Characterizing the hemodynamic response: effects of presentation rate, sampling procedure, and the possibility of ordering brain activity based on relative timing.

Rapid-presentation event-related functional MRI (ER-fMRI) allows neuroimaging methods based on hemodynamics to employ behavioral task paradigms typical of cognitive settings. However, the sluggishness of the hemodynamic response and its variance provide constraints on how ER-fMRI can be applied. In a series of two studies, estimates of the hemodynamic response in or near the primary visual and motor cortices were compared across various paradigms and sampling procedures to determine the limits of ER-fMRI procedures and, more generally, to describe the behavior of the hemodynamic response. The temporal profile of the hemodynamic response was estimated across overlapping events by solving a set of linear equations within the general linear model. No assumptions about the shape were made in solving the equations. Following estimation of the temporal profile, the amplitude and timing were modeled using a gamma function. Results indicated that (1) within a region, for a given subject, estimation of the hemodynamic response is extremely stable for both amplitude (r(2) = 0.98) and time to peak (r(2) = 0.95), from one series of measurements to the next, and slightly less stable for estimation of time to onset (r(2) = 0.60). (2) As the trial presentation rate changed (from those spaced 20 s apart to temporally overlapping trials), the hemodynamic response amplitude showed a small, but significant, decrease. Trial onsets spaced (on average) 5 s apart showed a 17-25% reduction in amplitude compared to those spaced 20 s apart. Power analysis indicated that the increased number of trials at fast rates outweighs this decrease in amplitude if statistically reliable response detection is the goal. (3) Knowledge of the amplitude and timing of the hemodynamic response in one region failed to predict those properties in another region, even for within-subject comparisons. (4) Across subjects, the amplitude of the response showed no significant correlation with timing of the response, for either time-to-onset or time-to-peak estimates. (5) The within-region stability of the response was sufficient to allow offsets in the timing of the response to be detected that were under a second, placing event-related fMRI methods in a position to answer questions about the change in relative timing between regions.

Comparison of brain activation during word retrieval done silently and aloud using fMRI.

Using functional MRI we compared the patterns of activation in an effortful word retrieval task (stem completion) performed both silently and aloud. The silent and overt conditions showed expected differences in activation magnitude in regions such as primary motor cortex. Some regions, such as frontal operculum and dorsolateral frontal cortex, showed similar activation magnitude across conditions. Thalamus was more active on the left in both conditions and showed a symmetric drop in activity in the silent compared with the overt condition. Putamen was also more active in the overt condition and showed a larger decrease in activity on the right than on the left in the silent compared with the overt condition. Thus it appears that silent and overt performance of this task engage the thalamus and putamen in different ways.

Voluntary orienting is dissociated from target detection in human posterior parietal cortex.

Human ability to attend to visual stimuli based on their spatial locations requires the parietal cortex. One hypothesis maintains that parietal cortex controls the voluntary orienting of attention toward a location of interest. Another hypothesis emphasizes its role in reorienting attention toward visual targets appearing at unattended locations. Here, using event-related functional magnetic resonance (ER-fMRI), we show that distinct parietal regions mediated these different attentional processes. Cortical activation occurred primarily in the intraparietal sulcus when a location was attended before visual-target presentation, but in the right temporoparietal junction when the target was detected, particularly at an unattended location.

The emotional modulation of cognitive processing: an fMRI study.

The functional neuroanatomy of visual processing of surface features of emotionally valenced pictorial stimuli was examined in normal human subjects using functional magnetic resonance imaging (fMRI). Pictorial stimuli were of two types: emotionally negative and neutral pictures. Task performance was slower for the negatively valenced than for the neutral pictures. Significant blood oxygen level dependent (BOLD) increases occurred in the medial and dorsolateral prefrontal cortex, midbrain, substantia innominata, and/or amygdala, and in the posterior cortical visual areas for both stimulus types. Increases were greater for the negatively valenced stimuli. While there was a small but significant BOLD decrease in the subgenual prefrontal cortex, which was larger in response to the negatively valenced pictures, there was an almost complete absence of other decreases prominently seen during the performance of demanding cognitive tasks [Shulman, G. L., Fiez, J. A., Corbetta, M., Buckner, R. L., Miezin, F. M., Raichle, M. E., & Petersen, S. E. (1997). Common blood flow changes across visual tasks: II. Decreases in cerebral cortex. Journal of Cognitive Neuroscience, 9, 648--663]. These results provide evidence that the emotional valence and arousing nature of stimuli used during the performance of an attention-demanding cognitive task are reflected in discernable, quantitative changes in the functional anatomy associated with task performance.

Areas involved in encoding and applying directional expectations to moving objects.

Two experiments used functional magnetic resonance imaging (fMRI) to examine the cortical areas involved in establishing an expectation about the direction of motion of an upcoming object and applying that expectation to the analysis of the object. In Experiment 1, subjects saw a stationary cue that either indicated the direction of motion of a subsequent test stimulus (directional cue) or provided no directional information (neutral cue). Their task was to detect the presence of coherent motion in the test stimulus. The stationary directional cue produced larger modulations than the neutral cue, with respect to a passive viewing baseline, both in motion-sensitive areas such as left MT+ and the anterior intraparietal sulcus, as well as motion-insensitive areas such as the posterior intraparietal sulcus and the junction of the left medial precentral sulcus and superior frontal sulcus. Experiment 2 used an event-related fMRI technique to separate signals during the cue period, in which the expectation was encoded and maintained, from signals during the subsequent test period, in which the expectation was applied to the test object. Cue period activations from a stationary, directional cue included many of the same motion-sensitive and -insensitive areas from Experiment 1 that produced directionally specific modulations. Prefrontal activations were not observed during the cue period, even though the stationary cue information had to be translated into a format appropriate for influencing motion detection, and this format was maintained for the duration of the cue period (approximately 5 sec).

Direct comparison of episodic encoding and retrieval of words: an event-related fMRI study.

Functional magnetic resonance imaging (fMRI) was used to compare directly episodic encoding and retrieval. During encoding, subjects studied visually presented words and reported via keypress whether each word represented a pleasant or unpleasant concept (intentional, deep encoding). During the retrieval phase, subjects indicated (via keypress) whether visually presented words had previously been studied. No reliable differences were found during the recognition phase for words that had been previously studied and those that had not been studied. Areas preferentially active during encoding (relative to retrieval) included left superior frontal cortex, medial frontal cortex, left superior temporal cortex, posterior cingulate, left parahippocampal gyrus, and left inferior frontal gyrus. Regions more active in retrieval than encoding included bilateral inferior parietal cortex, bilateral precuneus, right frontal polar cortex, right dorsolateral prefrontal cortex, and right inferior frontal/insular cortex.

A common network of functional areas for attention and eye movements.

Functional magnetic resonance imaging (fMRI) and surface-based representations of brain activity were used to compare the functional anatomy of two tasks, one involving covert shifts of attention to peripheral visual stimuli, the other involving both attentional and saccadic shifts to the same stimuli. Overlapping regional networks in parietal, frontal, and temporal lobes were active in both tasks. This anatomical overlap is consistent with the hypothesis that attentional and oculomotor processes are tightly integrated at the neural level.

Functional MRI studies of word-stem completion: reliability across laboratories and comparison to blood flow imaging with PET.

Functional magnetic resonance imaging (fMRI) based on blood oxygen level-dependent (BOLD) contrast has become an increasingly popular technique for mapping the brain. The relationship between BOLD-fMRI imaging and imaging of blood flow activation with positron emission tomography (PET) remains unclear. Moreover, BOLD imaging strategies and analysis procedures vary widely across laboratories. To examine the relationship between these different methods, we compared brain activation maps of a word-stem completion task obtained both using PET and using fMRI across two separate institutions (Washington University and Massachusetts General Hospital) with different acquisitions (gradient-refocused echo and asymmetric spin echo) and different analysis techniques. Overall, activation maps were highly similar across both fMRI methods and PET. A set of activated brain areas, in consistent locations in Talairach atlas space, were identified across all three studies, including visual striate and extrastriate, left prefrontal, supplementary motor area (SMA), and right cerebellar areas. Decreases in activation were also consistently observed in medial parietal, posterior insular, and medial inferior frontal areas. Some differences were noted that may be related to the silent performance of the task with fMRI. The largely consistent results suggest that comparisons can be made appropriately across imaging modalities and laboratory methods. A further implication of the consistencies, which extended to both increases and decreases in signal, is that the underlying brain physiology leading to BOLD contrast may be more similar to blood flow than originally appreciated.

Hemispheric specialization in human dorsal frontal cortex and medial temporal lobe for verbal and nonverbal memory encoding.

The involvement of dorsal frontal and medial temporal regions during the encoding of words, namable line-drawn objects, and unfamiliar faces was examined using functional magnetic resonance imaging (fMRI). Robust dorsal frontal activations were observed in each instance, but lateralization was strongly dependent on the materials being encoded. Encoding of words produced left-lateralized dorsal frontal activation, whereas encoding of unfamiliar faces produced homologous right-lateralized activation. Encoding of namable objects, which are amenable to both verbal and nonverbal encoding, yielded bilateral dorsal frontal activation. A similar pattern of results was observed in the medial temporal lobe. These results indicate that regions in both hemispheres underlie human long-term memory encoding, and these regions can be engaged differentially according to the nature of the material being encoded.

Model-based scatter correction for fully 3D PET.

A method is presented that directly calculates the mean number of scattered coincidences in data acquired with fully 3D positron emission tomography (PET). This method uses a transmission scan, an emission scan, the physics of Compton scatter, and a mathematical model of the scanner in a forward calculation of the number of events for which one photon has undergone a single Compton interaction. The distribution of events for which multiple Compton interactions have occurred is modelled as a linear transformation of the single-scatter distribution. Computational efficiency is achieved by sampling at rates no higher than those required by the scatter distribution and by implementing the algorithm using look-up tables. Evaluation studies in phantoms with large scatter fractions show that the method yields images with quantitative accuracy equivalent to that of slice-collimated PET in clinically useful times.

Tracer-kinetic analysis for measuring regional cerebral blood flow by dynamic nuclear magnetic resonance imaging.

Measurement of regional cerebral blood flow in vivo has proved useful in the study of normal and diseased states in the brain. This circumstance has led to a variety of techniques for its quantitative determination and has continued to motivate the search for ever safer and more accurate methods of measurement. Recently, the use of nuclear magnetic resonance (NMR) in medical imaging has stimulated efforts to make it the basis for a non-invasive method of measuring flow in the brain. New advances in fast NMR imaging (MRI) provide data potentially amenable to analysis by tracer-kinetic methods. Such an analysis has not previously been available. In this paper we present theoretical results that may permit measurement of brain blood flow by NMR. The data interpreted by our model are those generated by a novel MRI protocol developed by Perman et al. (1992, Magn. Reson. Med. 28, 74-83; Radiology 185(P), Abstr. 154, 127) that is entirely compatible with existing routine MRI procedures. These data are fast dynamic NMR signals that reflect passage of an intravenously administered paramagnetic contrast agent serving as a plasma tracer. Our equations show how to use such data sequences to determine plasma mean transit time, plasma volume, and plasma and whole-blood flow in arbitrarily selected regions of interest in the brain. The theory accounts rigorously for recirculation of tracer to the imaged regions. Our analysis provides an explanation for the linear relationship observed experimentally by others between regional vascular volumes and time integrals of vascular-tracer residue curves, and shows that this relationship remains valid in the presence of tracer recirculation.

Maximum-likelihood reconstruction of transmission images in emission computed tomography via the EM algorithm.

The expectation-maximization (EM) algorithm for computing maximum-likelihood estimates of transmission images in positron-emission tomography (PET) (see K. Lange and R. Carson, J. Comput. Assist. Tomogr., vol.8, no.2, p.306-16, 1984) is extended to include measurement error, accidental coincidences and Compton scatter. A method for accomplishing the maximization step using one step of Newton's method is proposed. The algorithm is regularized with the method of sieves. Evaluations using both Monte Carlo simulations and phantom studies on the Siemens 953B scanner suggest that the algorithm yields unbiased images with significantly lower variances than filtered-backprojection when the images are reconstructed to the intrinsic resolution. Large features in the images converge in under 200 iterations while the smallest features required up to 2,000 iterations. All but the smallest features in typical transmission scans converge in approximately 250 iterations. The initial implementation of the algorithm requires 50 sec per iteration on a DECStation 5000.