| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review |
1 Both authors: Department of Radiology, Miami Children's Hospital, 3100 62nd Ave., Miami, FL 33155-3009.
Received April 3, 2000;
accepted after revision October 6, 2000.
Address correspondence to N. R. Altman.
Introduction
 Top Introduction Auditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
Currently, functional imaging of cerebral activity relies on coupling of blood flow to neuronal activity and metabolism with new MR imaging techniques that permit rapid and repetitive acquisition of images and good temporal and spatial resolution. Several procedures have been used to show the direct relationship between neural metabolism and blood supply. Early positron emission tomography (PET) indicated neuronal activity and metabolism were accompanied by local changes in cerebral blood flow [1], cerebral blood volume [2], and blood oxygenation [2, 3]. In 1990, Belliveau et al. [4] showed that changes in regional cerebral blood volume resulting from functional visual activation could be localized in the occipital cortex with MR imaging. Using a bolus injection of a paramagnetic contrast medum, they showed that the transit curve of contrast medium could be fitted to the intensities of a series of images taken during the contrast injection. This procedure allowed an estimation of blood volume by contrast enhancement. Differences in computed cerebral blood volume images obtained during activated and nonactivated states identified the location of blood volume changes resulting from functional activity.
Initial functional MR (fMR) imaging was based on T1-weighted images. These sequences depended on saturation of the signal within a given volume. Cortical activation increased inflow of blood into the underlying voxel, producing a volume of unsaturated protons causing a shortening of apparent T1. The advent of echoplanar imaging used the T2* effect based on magnetic susceptibility, which led to signal loss due to local field inhomogeneities of deoxyhemoglobin. In 1982, it was established that local susceptibility changes resulted from the intrinsic paramagnetism of deoxyhemoglobin [5]. The main contribution to further development of fMR imaging came from Ogawa et al. [6] and Ogawa and Lee [7], who showed that MR imaging could be used to document the regional changes in brain oxygenation.
Although susceptibility effects are reduced at clinical imaging field strengths, they are sufficient to enable the localization of functional activation by means of hardware and software developed in the last 5-10 years.
Signal production in fMR imaging is dependent on the paramagnetic effect of deoxyhemoglobin. Focal brain activation results in regional increases of cerebral blood volume, cerebral blood flow, and oxygen delivery, with a modest increase in oxygen extraction [3]. The ratio between oxyhemoglobin and deoxyhemoglobin of venous blood increases and results in less tissue—blood susceptibility difference. Finally, less susceptibility results in less intravoxel dephasing that, in turn, leads to more signal observable on T2* and T2 images. This result is termed the blood oxygenation level—dependent effect. Although the direct cause of the increase in the signal is the reduced paramagnetic effect of the deoxyhemoglobin, more oxyhemoglobin results in greater signal, which is also directly related to amount of cerebral blood flow, cerebral blood volume, and neural activity.
These small signal changes occur in a range no greater than 5-7% at 1.5 T [5, 8]. Functional responses to diverse types of stimuli have been obtained from mechanical, motor, sensory, auditory, visual, gustatory, verbal, and sophisticated cognitive tasks. Results previously obtained on PET have been reproduced on fMR imaging. Most work is directed toward research in brain mapping, but studies in clinical and diagnostic fields are increasing. Recent studies have focused on characterization of the underlying mechanisms of learning disabilities, a vast group of entities in which there are no anatomic abnormalities identified to date. fMR imaging studies of patients with dyslexia [9] and attention deficit—hyperactivity disorder [10] have been reported. In addition, fMR imaging has proved to be a good alternative to Wada's test in lateralizing language [11, 12]. fMR imaging can provide an anatomically accurate topographic map of the functional cortex and can allow useful presurgical planning [13, 14].
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
|
There is a tonotopic array in the human primary auditory area that has been shown by different techniques including magnetoencephalography [18], PET [19], and fMR imaging [20]. These studies have shown that tones have a posteromedial-to-anterolateral organization corresponding to high-to-low frequencies, respectively. Other studies have found medial-to-lateral distribution related to low-to-high frequencies [21].
The primary auditory cortex projects to several adjacent association areas in the superior and middle temporal gyri, including Brodman's areas 22 and 39. Connections with the prefrontal cortex, multimodal areas, and limbic regions have been found in monkeys and humans [22] and are probably related to the context of the stimulus.
Auditory pathways have shown some functional brain asymmetry. Dichotic listening tasks, in which different auditory stimuli are presented simultaneously to both ears, historically revealed consistent right ear dominance in righthanded subjects for hearing verbal material such as digits, words, and syllables [23, 24]. Conversely, a left ear dominance has been obtained for the recognition of nonverbal sounds [23, 25]. Since these findings were published, many other experiments have shown similar laterality for speech performance. These tests include event-related evoked potentials [26], PET [27], intracarotid amobarbital testing [28], and fMR imaging [12]. From these findings, it seems that the left temporal lobe is better at processing language information, whereas the right temporal lobe has the advantage in processing natural sounds without verbal meaning.
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
A paradigm consists of an experiment that uses a particular task to elicit a specific brain function. To be reliable, the task must be simple, precise, reproducible, and easy to perform without the subject moving. The active task must be compared with a resting or reference task. The active tasks may be produced by a sequence of repeated stimuli (ON epoch) compared with a resting or reference task (OFF epoch) in the block technique. The event-related technique uses a single stimulus (the event) that defines the active task.
The auditory paradigm is a demanding paradigm because the OFF state of the auditory examination is hardly a resting condition. The noise generated from switching the gradients of the echoplanar imaging sequence produces activation of the auditory area [29, 30]. Initially the sound level of this sequence was measured by Cho et al. [31] at 100 dB or more with peaks at 500 Hz. More recently, peak noise levels generated by the echoplanar imaging beep have been found ranging from 117 dB on a 1.5-T to 133 dB on a 3.0-T imager [32]. The echoplanar imaging—gradient sound is repetitive and constant throughout the ON and OFF epochs. The postprocessing cross-correlation analysis should remove the effects of any constant auditory stimulus.
The noise from the MR scanner may be reduced by noise cancellation devices; by earplugs or headphones with insulated conducting tubes; or by synchronizing the stimuli with the pulses of the MR sequences. Some variations of common MR techniques have been proposed to reduce the effect of the background noise. Synchronizing auditory stimuli and fMR imaging gradient pulses allows presenting the auditory stimulus in blocks preceding the gradient switching. With this strategy, it is possible to get long silent intervals after the echoplanar imaging—gradient acoustic noise [33,34,35]. A similar technique proposed by other authors [21, 36, 37] is timing the gradient pulses with task, using the physiologic delay between neuronal activity and its hemodynamic response. Also, extremely long TR has been used by different authors trying to leave a longer time of silence between switches. Robson et al. [38] used a 20-sec TR, and Stippich et al. [39] used a 15-sec TR with good results.
Another possibility is to present the stimulus with frequencies different from the noise generated by the MR gradients. Each frequency activates a specific portion in Heschl's gyrus. The frequencies of the gradients are grouped at 200 and 2500 Hz [40]. Better activation may be obtained with stimulus of different frequencies. Currently, the best approach appears to have the frequency of the stimulus far from the scanner noise frequencies. The use of earphones and computer postprocessing also allows for cancellation of the background acoustic noise.
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
|
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
External Factors
Effects of background noise.—The acoustic noise of the
echoplanar imaging pulse sequence consists of frequencies grouped around 200
and 2500 Hz at high intensity (volume). This stimulus produces activation of
the primary auditory cortex relative to a baseline of virtual silence
[41] or contrasted with a real
silence condition in an elegant experiment conducted by Bandettini et al.
[42] and replicated by Ulmer
et al. [43]. Optimizing
external and internal factors made it possible to get activation related to a
given sound stimulus because the cortex can extract the foreground stimulus
from the background and the background noise can be subtracted by computer
postprocessing. Scheich et al.
[44] showed that there is an
intrinsic capability of the auditory cortex to discriminate the foreground
targeted stimulus from the background stimulus. The area involved is in the
rostral aspect of the Heschl's gyrus. Background acoustic scanner noise
increases the pure tone cortical thresholds, particularly in the frequency
range of 125-8000 Hz, although without a uniform effect across the frequency
spectrum [30].
Effects of intensity.—The extent of activation in the superior temporal gyrus increases with the stimulus intensity [45, 46]. The experiment of Jäncke et al. [46] in 14 subjects who received stimuli at 75, 85, and 95 dB showed larger activation with increasing intensity. These findings also revealed that reducing the stimulus intensity to lower than 60 dB would make activation impossible to detect.
Effects of pitch (frequency).—It is important to consider the frequency of the tones presented in fMR imaging. Many frequencies have been used successfully. Authors have used pure-tone stimulus at different frequencies: 55 and 880 Hz [20]; 500, 700, 1000, 1400, 2000, 2800, and 4000 Hz [47]; 440 Hz [38]; 500 Hz [48]; 200, 400, 600, 800, 1000, and 2000 Hz [46]; and 1000 and 4000 Hz [49]. Scheich et al. [44] used pure tones given randomly at 11 different frequencies between 3 and 4.5 kHz. Pugh et al. [50] used falling and rising tones between 1200 and 1900 Hz. All these researchers obtained activation using these frequencies. Activation is greater with a stimulus at 1000 Hz rather than at 4000 Hz and with a progression of frequencies (stepped tones) versus a single one (pure tone) [45].
Effects of stimulus rate and duration.—The rate of the stimulus directly correlates with the metabolic rate and blood oxygenation level—dependent signal intensity. Neuronal recovery and blood flow regulations may alter this relationship. To our knowledge, there is no information about the effects of the rate of auditory stimulus with pure tones. Two experiments with speech tasks have targeted this problem. Because the words were used as pure auditory stimulus and no cognitive demands were asked of the subjects, these experiments may be considered pure auditory stimuli. In the first experiment, syllables were presented to five healthy subjects at rates ranging from 0.17 to 2.5 Hz. The magnitude of the activation in the superior temporal lobe increased in a nonlinear manner with increasing stimulus rate [51]. In the second experiment, English nouns were presented at rates ranging from zero (rest) to 130 words per minute. In this study, there was an increase of activation in the superior temporal gyrus and transverse temporal sulcus with increasing rate of word presentation, peaking at 90 words per minute [52].
Effects of tones, chords, music, and other kinds of stimulus.—Most of the work in auditory fMR imaging has been performed with tones, syllables, and words because the more basic the function investigated, the more intersubject reproducibility is expected. Other stimuli have also been used to evoke activation of the auditory system: these include a variety of sounds and tones played by different musical instruments, environmental sounds, white noise, and even a mother's or father's voice in sedated children [53, 54]. Music has also been used as a stimulus for the auditory area [29, 34, 55], obtaining the largest activation due to the effect on secondary associative areas involved in memory, feelings, and enjoyment. In our laboratory, music produced better activation of the primary auditory area compared with tones. An example is presented in Figure 2.
|
Internal Factors
Habituation.—Habituation can be defined as the reduction of
activation of the cortex after prolonged exposure to a stimulus. It is unknown
whether the mechanism of habituation resides in the primary auditory area,
subcortical structures, or higher cognitive levels. In PET experiments, six
healthy subjects were examined to test the effect of continuous white noise
presentation ranging from 10 to 120 sec. Habituation was seen in the thalami,
but no significant cerebral blood flow reduction between the early and later
phases of the noise stimulation was observed in the primary auditory areas.
Because most works have tailored the ON epochs to about 30 sec, no significant
habituation to the task would be expected.
Lag of response.—Generally, there is a finite time between a stimulus and the activation of the cortex. For visual, motor, and sensory stimulus, this activation peaks at 5-10 sec [56, 57]. The auditory-lag response appears to be slightly longer with a peak at 10.5 sec. Imaging acquisition and task paradigms can be constructed to take advantage of this lag of response.
A hypothetic example consists of acquiring the fMR images at the end of the 10.5-sec lag time after a 1-sec auditory stimulus. This lag allows the cortical activation to be uncorrupted by the scanner noise during the stimulus presentation. A similar model was used by Yang et al. [21].
Tonotopy.—The auditory system has a tonotopic organization throughout all its levels. This organization is present in the auditory receptors at the cochlea, the ascending pathways, and the cortex [15]. The tonotopy organization of Heschl's gyrus was shown on fMR imaging by Wessinger et al. [20] in 1997. This group found, in five of six subjects (two left-, three right-handed), that this tonotopy occurs in an asymmetric fashion, more pronounced in the left hemisphere. High-to-low frequency representation courses in a posteromedial-to-anterolateral direction. A second study conducted by Bilecen et al. [58] found a slight craniocaudal shift for the higher frequencies, more pronounced in the right temporal lobe. Bass tones appear to activate larger areas than higher tones. The relationship between the asymmetry of the tonotopic organization and the capability to encode language has not been established.
Contralaterality and Lateralization
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
The human auditory cortex receives input from both ears, however predominantly from the contralateral one. The degree of this contralaterality is given for the amount of fibers that cross the midline at the cochlear nuclei, the superior olivary nuclei, and the inferior collicular level. Quantification of contralaterality has been assessed with neurophysiologic procedures. This is reported with ratios to describe the amount of activation of one ear compared with the other. For example, if the right ear is the contralateral one, then the sound received would be 66.6% to the left hemisphere and 33.3% to the right hemisphere or a ratio of 2:1. Monaural studies with evoked potentials [59] and magnetoencephalography [60] have shown contralateral predominance in ratios ranging from 1.15:1 to 1.3:1. With fMR imaging with monaural presentation, the degree of contralateral activation was found in about 2:1 in one work [47] and in ratios of 3.4:1 for the right ear and up to 5.2:1 for the left ear in the other study [61].
The lateralization of auditory-language function to the left hemisphere is well known and is not the theme of this review. Lateralization is also expressed in ratios because one hemisphere is rarely 100% dominant. Lateralization for basic nonverbal stimulus has also been reported. The initial publications using PET found a right-sided dominance in agreement with lesional models. Music appreciation comparing listening to melodies versus listening to noise resulted in cerebral blood flow increases in the right superior temporal and right occipital cortices. Evaluation of pitch resulted in right frontal-lobe activation. Analysis of memory for notes revealed the participation of the right frontal, temporal, and parietal lobes and the insular cortex. With activation on the right side, the cerebral blood flow decreased in the left primary auditory cortex [62, 63]. The first study using fMR imaging reported by Binder et al. [51] found no lateralization in right-handed subjects with consonant—vowel meaningless syllables (speech without language stimulus). More recent studies with PET and fMR imaging have found activation preferentially in the left hemisphere. With PET, left-sided dominance was found for familiarity, pitch, and rhythm; right-sided dominance was seen only for timbre [64]. Millen et al. [49], using pure-tone stimulus, found bilateral activation with lateralization in the range of 2.75-1 in the left superior—temporal gyrus. Scheffler et al. [61] found slight left hemispheric dominance (left temporal) for binaural stimulation in eight of 10 right-handed volunteers with a pure-tone stimulus of 1000 Hz, presented with a rate of 6 Hz. Similar findings were reported by Strainer et al. [45], Bilecen et al. [65], and Formisano et al. [66], using tones of 500 and 4000 Hz. It has been found that the left posterosuperior—temporal gyrus is implicated in the detection of acoustic changes, not only in speech but also in nonspeech stimuli [67].
The predominance of activation in the left superior—temporal gyrus is a challenging finding because the lesional and neuropsychologic models have suggested that the right hemisphere is dominant for nonverbal processes. It appears that for nonspeech auditory stimuli and music, there is a cascade of events that occurs in the brain that may involve both hemispheres. More research with functional studies and clinical lesional cases is necessary to elucidate this complex cascade and to determine precisely how each hemisphere contributes in music cognition.
Attention and Memory as Cognitive Confounds
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
Different methods have been used to evaluate attention in clinical auditory tasks. Activation in the posteroparietal, superior temporal, and inferior frontal regions increases when listeners perform tasks with dichotic conditions, during which greater demands are made on auditory selective attention [50]. The activation of primary and secondary auditory areas is greater in dual-task situations in which visual and auditory stimuli are concurrent and discrimination is needed [46]. A distinction should be made between attention and detection conditions. The attention condition requires attentive listening to stimuli, whereas a detection condition requires detection of a specific target, adding another process. The latter results in a stronger activation [46].
In addition to attention and detection of a given stimulus, a subject can be asked to memorize commands to perform during the task. Activation will increase as a result of the memory requirements. This type of memory is called working memory. Activation may be seen in the lateral and medial prefrontal cortex, temporal lobe (including insula and hippocampus), parietal—occipital cortex, congulate, thalamus, and superior colliculus [68]. Primary auditory areas are involved in memory and retrieval of melodies [55]. Tasks intended to explore the primary auditory area must avoid stimuli that can be part of the repertoire of the subject's experience.
Potential Clinical Applications of Auditory fMR Imaging
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
For monaural deaf subjects, the lateralization ratio is reduced compared with binaural responses of normal-hearing subjects. This difference indicates plasticity or reorganization of auditory pathways [61]. Activation in the auditory cortex during cochlear electric stimulation in deaf patients has been shown [69]. This finding may make fMR imaging studies suitable to determine cochlear implant candidates.
Patients with lateralized tinnitus showed absent or reduced activation on fMR studies in the contralateral inferior colliculus [70]. The differences of activation observed in groups of patients with different types of tinnitus suggest a correlation with prognosis. The ability to classify tinnitus by anatomic and functional methods allows another way to evaluate therapeutic results.
In one patient, fMR imaging has also been used successfully to evaluate the time course of compensatory cortical reorganization of auditory centers after unilateral destruction of the cochlear nerve [71]. Further investigation is needed to determine if these findings might be used to evaluate the outcome in patients with unilateral deafness. Patients suffering from chronic fatigue syndrome show less lateralization in the superior temporal and angular gyrus than that found in healthy subjects performing an auditory working memory test.
Schizophrenia seems to be related to a reduced left and increased right temporal—cortical activation to auditory perception of speech [72]. fMR imaging has shown the involvement of the primary auditory areas in psychotic auditory hallucinations [73].
Mapping comprehension to verbal material in patients with left temporal tumors has shown the spatial distribution of activation differs from the pattern observed in control subjects [74]. Similar changes could theoretically appear with tones, chords, or music.
Activation of Heschl's gyrus and related areas obtained using a mother's or father's voice in sedated children may be produced by a primary auditory stimulus because the activation was found in preverbal infants. This finding opens a promising field for further investigations of language skills and disorders in children.
In conclusion, fMR imaging is based on changes of the oxy- and deoxyhemoglobin ratio from blood flow and volume in areas in which neuronal activity increases metabolism. Passive experiments have shown activation and tonotopic organization of primary auditory areas with basic tone stimuli, despite the effect of persistent background noise from the MR scanner and cognitive confounding variables. Although fMR imaging is still evolving, this tool is reliable and easy to use. Clinical applications are on the horizon. Presently, preoperative mapping is being increasingly used in many centers throughout the world. Basic verbal functions related to the primary auditory system and of language development in the pediatric population is not well understood but will be explored by fMR imaging. Auditory stimuli present the possibility of passive paradigms in the fMR imaging workup, which may have significant value in the examination of children and comatose and sedated patients. Hemispheric organization, music dominance, nonverbal memory, and auditory discrimination will be elucidated with auditory functional brain mapping with fMR imaging.
Acknowledgments
We thank Susan DeBusk for the work done in manuscript and reference
production.
|
TopIntroductionAuditory BackgroundMethodologic IssuesSequence SettingsVariablesContralaterality and...Attention and Memory as...Potential Clinical Applications...References |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. H. Hoon JR, K. M. Belsito, and L. M. Nagae-Poetscher Neuroimaging in Spasticity and Movement Disorders J Child Neurol, January 1, 2003; 18(1_suppl): S25 - S39. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
