摘要：在临床应用方面，fMRI协助神经科医生在大脑的手术前的功能定位，揭示神经病的病因，协助神经病医生研究异常的大脑功能等方面发挥重要作用。本文在fMRI概念，技术背景和各种应用方面综述。fMRI 的不足之处及未来的发展也在文章进行了讨论。

CLC number: R445.2Document code: AArticle ID: 1005-202X(2000)01-0007-01

Functional MRI in neuroscience

—Part II: a review in application

Charlene X.Tan

（Laboratory of Diagnostic Radiology Research, National Institute of Health, Room B1N256）

Abstract: In the clinical environment, fMRI should aid neurosurgeons in pre-surgical diagnostics of the brain, assist neurologists in unraveling neurological hypotheses, and help psychiatrists in studying the abnormal brain functions .The overviews were given for fMRI concepts, technique backgrounds and various applications. Certain limitations and possible future developments have been discussed.

III. Applications

Comparing to other neuroimging techniques such as PET, SPECT, MEG, EEG, fMRI is noninvasive, flexible and anatomical imaging obtainable for the same subject at the same fMRI are more attractive for most neuroscientists and clinicians. There are three major categories of brain fMRI applications: mapping of cortical functional structures such as auditory cortex, visual cortex and motor cortex; studying the association and connectivity of cognitive and behavioral information processing; clinical applications including neurological patient recovery, brain surgery planning, mental illness patient study and screening. Figure 8and 9 demonstrate fMRI obtained via a block-stimulus design mapped on to the individual′s anatomic image. Figure 8 is a fMRI map of five transverse slices, obtained with the subject (Dutch woman) listened different types of music during the activation periods .The music pieces are violin (a music instrument with strings) performance, symphony orchestra (many different types of classic music instruments), rock and roll (heavy music instruments), and kinder music with clear words in Dutch .The increased activation areas in the temporal lobe by the music bandwidth and by the words are clearly visible, Figure 9 is a fMRI map of three orthogonal projections (transverse,sagittal and coronal),obtained while subject performing mental arithmetic during the activation periods . the three major activation foci are front lobe (frontal eye field ),frontal lobe (attention )and occipital lobe (reading the mathematics formula).

fig.8fMRI maps (black) obtained with the subject (a Dutch woman) listened different types of music during the activation periods .The music pieces presented to the subject are produced by different instruments: violin (a music instrument with strings) performance, symphony orchestra (many different types of classic music instruments), rock and roll (heavy music instruments), and kinder music with clear words in Dutch .The increased activation areas in the temporal lobe by the music bandwidth and by the words are clearly visible.

fig.9fMRI map of three orthogonal projections (transverse, sagittal and coronal) obtained while subject performing mental arithmetic during the activation periods.Auditory

Because of the background noise generated by the rapid current ramping in the MRI coils, auditory stimuli is the most difficulty one to design and to apply among the three major stimuli. Earlier auditory fMRI was mainly in studying tonotopic organization of the primary auditory cortex^［1,2］. Binder et al.^［3］ demonstrated bilateral superior temporal lobe MRI signal increase that were coincident with auditory stimulus presentation and performance of the task. The magnitude of this response increased in a monotonic, non-linear manner with increasing stimulus rate. This rate-response relationship was nearly identical in right and left hemispheres. The relationship may reflect metabolic activity integrated over time and subject to non-linear characteristics of neuronal recovery or blood flow regulation. The same group also studied fMRI with auditory stimuli including non-speech noise, meaningless speech sounds, single words, and narrative text^［2,3］. Signal changes in the superior temporal gyrus and superior temporal sulcus were observed bilaterally. Speech stimuli were associated with significantly more widespread signal changes than was the noise stimulus, while no consistent differences were observed between responses to different speech stimuli.Alater report by Millen et al.^［4］ stated that activation in both text and pure-tone presentation did not vary with the intensity of the auditory stimulus and elicited a dominant response in the left temporal lobe (with text presentation )^［5,6］.Not until recent, increasing amount of reports in auditory fMRI become available .Guimaraes et al. Of MGH^［7］ study the brain stem fMRI to find how humans process auditory listening information by reducing the cardiac and pulsatile motion. Scheich et al.^［8］ using a new type of earphones to reduce the scanner noise and to apply the auditory stimuli. They found that the auditory foreground-background decomposition involves a specialized non-primary auditory cortex field, the auditory pattern recognition process, which combines simultaneous and sequential grouping in complex sound sequences, activates topographically distinct areas of human auditory cortex. Schlosser et al.^［9］ compared the activation location of auditory language comprehension with pure tone. Language testing produce strong activation within the left superior temporal sulcus (the dominant temporal lobe). Lesser activation was seen in homotopic right hemisphere locations with non-native language (tone).

Visual

Current fMRI application using visual stimuli has been advanced to a much complex and mature level in comparison with auditory fMRI. Culham et al.^［10］ used fMRI to investigate cortical regions involved in attentive tracking of moving items with multiple targets. They found that attentive tracking is mediated by a network of areas that includes parietal and frontal regions responsible for attention shifts and eye movements and the MT complex, thought to be responsible for motion perception .He et al.^［11］ convinced that the area of the brain termed MT/V5 (+) which includes the middle temporal visual area MT/V5 along with adjacent motion-sensitive areas such as MST——is correlates with the strength of perceived motion aftereffect (MAE). Beason-Held et al.^［12］ studied visual field in elementary form processing. They find both random and correlated texture stimulation resulted in increased signal intensity primarily in the striate cortex, with slight involvement of the cuneus and middle occipital, lingual and fusiformm gyri. Correlated texture stimulation resulted in activation of significantly greater, regional extent than that produced by random textures. Correlated stimulation additionally resulted in middle temporal activation. The increase in regional activation during the correlated condition suggests that increased recruitment of neuronal populations occur in response to textures containing visually salient features. This increased recruitment occurs within striate, extrastriate a