Functional transcranial Doppler ultrasound complements other functional imaging modalities, with its high temporal resolution measurement of stimulus-induced changes in cerebral blood flow within the basal cerebral arteries. This Methods paper gives step-by-step instructions for using functional transcranial Doppler ultrasound to perform a functional imaging experiment.
Functional transcranial Doppler ultrasound (fTCD) is the use of transcranial Doppler ultrasound (TCD) to study neural activation occurring during stimuli such as physical movement, activation of tactile sensors in the skin, and viewing images. Neural activation is inferred from an increase in the cerebral blood flow velocity (CBFV) supplying the region of the brain involved in processing sensory input. For example, viewing bright light causes increased neural activity in the occipital lobe of the cerebral cortex, leading to increased blood flow in the posterior cerebral artery, which supplies the occipital lobe. In fTCD, changes in CBFV are used to estimate changes in cerebral blood flow (CBF).
With its high temporal resolution measurement of blood flow velocities in the major cerebral arteries, fTCD complements other established functional imaging techniques. The goal of this Methods paper is to give step-by-step instructions for using fTCD to perform a functional imaging experiment. First, the basic steps for identifying the middle cerebral artery (MCA) and optimizing the signal will be described. Next, placement of a fixation device for holding the TCD probe in place during the experiment will be described. Finally, the breath-holding experiment, which is a specific example of a functional imaging experiment using fTCD, will be demonstrated.
In neuroscience research, it is often desirable to monitor real-time brain activity noninvasively in a variety of environments. However, conventional functional neuroimaging modalities have limitations that impede the ability to capture localized and/or rapid activity changes. The true (non-jittered, non-retrospective) temporal resolution of functional magnetic resonance imaging (fMRI) is currently of the order of a few seconds1, which may not capture transient hemodynamic changes linked to transient neural activation. In another example, although functional near-infrared spectroscopy (fNIRS) has high temporal resolution (milliseconds) and reasonable spatial resolution, it can only probe hemodynamic changes within the cerebral cortex and cannot provide information about changes taking place in the larger arteries supplying the brain.
In contrast, fTCD—classified as a neuroimaging modality—“imaging” refers to the dimensions of time and space, rather than two orthogonal spatial directions that are more familiar in an “image”. fTCD provides complementary information to other neuroimaging modalities by measuring high temporal resolution (typically 10 ms) hemodynamic changes at precise locations within vessels of the basal cerebral circulation. As with other neuroimaging modalities, fTCD may be used for a variety of experiments such as studying lateralization of cerebral activation during language-related tasks2,3,4, studying neural activation in response to various somatosensory stimuli5, and exploring neural activation in various cognitive stimuli such as visual tasks6, mental tasks7, and even tool production8.
Although fTCD offers several advantages for use in functional imaging, including low cost of equipment, portability, and enhanced safety (compared to Wada test3 or positron emission tomography [PET] scans), operation of a TCD machine requires skills obtained by practice. Some of these skills, which must be learned by a TCD operator, include the ability to identify various cerebral arteries and the motor skills necessary to precisely manipulate the ultrasound probe during the search for the relevant artery. The goal of this Methods paper is to present a technique for using fTCD to perform a functional imaging experiment. First, the basic steps for identifying and optimizing the signal from the MCA, which perfuses 80% of the cerebral hemisphere9, will be listed. Next, placement of a fixation device for holding the TCD probe in place during the experiment will be described. Finally, the breath-holding experiment, which is one example of a functional imaging experiment using fTCD, will be described, and representative results will be shown.
All human subject research was performed in accordance with the Institutional Review Board of the University of Nebraska-Lincoln, and informed consent was obtained from all subjects.
1. Locating the MCA signal by freehand TCD
NOTE: “Freehand” TCD refers to operation of TCD with a handheld transducer to find a CBFV signal before beginning an fTCD experiment.
Figure 1: Representation of the circle of Willis and the major arteries of the cerebral circulatory system. The bifurcation of the ICA into the ACA and MCA is marked with a black circle. The M1 segment of the MCA is shown. This figure has been modified from24. Abbreviations: ACA = anterior cerebral artery; Bif. = bifurcation; ICA = internal carotid artery; MCA = middle cerebral artery. Please click here to view a larger version of this figure.
Figure 2: The transtemporal window (marked by the dashed ellipse), zygomatic arch (arrow), and subwindows11. (A) Frontal subwindow. (B) Anterior subwindow. (C) Middle subwindow. (D) Posterior subwindow. Please click here to view a larger version of this figure.
Figure 3: Sample Doppler spectra and M-mode images from midpoint of M1 segment of the MCA. (A) Spectrum taken right after applying transducer to the temporal window, just in front of the ear. (B) Sample Doppler spectrum at same location and depth as (A). The only change is that the transducer has been angled upwards (superiorly) slightly. In both (A) and (B), depth = 50 mm, gain = 50, sample volume = 12 mm, power = 420 mW/cm2, and filter = 100 Hz. Please click here to view a larger version of this figure.
Figure 4: Spectral Doppler (top) and M-mode (bottom) image of bifurcation of the ICA into the MCA and ACA. Depth = 65 mm, gain = 50, sample volume = 12 mm, power = 420 mW/cm2, and filter = 100 Hz. Please click here to view a larger version of this figure.
2. Relocating the MCA after placing a fixation device
NOTE: For fTCD experiments, it is necessary to monitor CBFV for 10–90 min or longer. Therefore, a fixation device (Figure 5) is crucial to provide stability.
Figure 5: Subject wearing custom fixation device. Please click here to view a larger version of this figure.
3. Performing a breath-hold maneuver
NOTE: This section is given as an example of a functional experiment that may be performed using the experimental setup described in section 1 and section 2.
Figure 3 shows sample Doppler spectra and color M-modes from the midpoint of the M1 segment of the MCA. Figure 3A,B were taken at the same position on the scalp, but at different angles. Note how a very small change in angle, without changing the contact position on the scalp, can greatly improve Doppler signal strength, as shown by the higher-intensity yellow coloring of the spectrogram in Figure 3B. Note also that the M-mode in Figure 3B shows two arteries (blue and red, corresponding to the ACA and MCA, respectively).
Figure 4 shows a sample Doppler spectrum and M-mode from the bifurcation of the ICA into the ACA and MCA. Note the overlapping red- and blue-shaded regions in the M-mode image denoting the MCA and ACA, respectively. Also note the symmetry of the Doppler spectral waveform when comparing flow towards the transducer (positive) with flow away from the transducer (negative).
Figure 6 shows sample spectra and M-mode images from different time points in the breath-hold maneuver. Figure 6A shows the baseline TCD spectrum and M-mode at the beginning of breath-holding. Note the mean velocity of 56 cm/s. Figure 6B shows the TCD spectrum and M-mode at the end of breath-holding. Note that the mean velocity has now increased to 70 cm/s. Figure 6C shows the TCD spectrum and M-mode after the end of breath-holding. Note the undershoot in velocity below baseline values, with the mean dropping to 47 cm/s. Note that the ACA is visible as flow away from the transducer in the Doppler spectra.
Figure 7 shows the entire breath-holding experiment. Note that the envelope remains elevated for approximately 15 s after breath-holding ends, falls to values lower than those at the beginning of breath-holding for ~20 s, and then finally recovers to baseline values. Note that the ACA is visible as flow away from the transducer in the Doppler spectrum.
Figure 6 and Figure 7 display good signal intensity in the MCA portion of the TCD spectrum (the MCA is represented by the positive velocities); note how the white line which represents the envelope follows the TCD spectrum very accurately when the spectrum is bright. The spectra of Figure 6 and Figure 7 could be improved by decreasing the monitoring depth by 5–10 mm so that the ACA portion of the TCD spectrum would not be visible (the ACA is represented by negative velocities) and by changing the scale of the vertical axis in the TCD spectrum to run from approximately -100 cm/s to 100 cm/s, which would allow maximum velocity sampling of the TCD spectrum in the vertical direction.
Figure 8 shows examples of bilateral TCD spectra and M-modes suitable for bilateral fTCD. Figure 8A and Figure 8B demonstrate acceptable, but not optimal, bilateral spectra and M-modes. Note how the gain is higher in Figure 8A (left MCA) than in Figure 8B (right MCA) to compensate for the weaker signal, and how the envelope quality in Figure 8A is slightly poorer than in Figure 8B. Also note how the maximum velocity at systole in Figure 8A is slightly lower than in Figure 8B. By contrast, note how the two spectra in Figure 8C and Figure 8D are very similar in terms of settings, including depth, gain, power, and sample volume, and how the spectral waveforms on both sides have similar maximum velocities and shapes. To address this, it is recommended that the spectrum from the left MCA be consistently placed in the left window and the spectrum from the right MCA in the right window, especially for experiments involving lateralization of blood flow.
Figure 6: Sample Doppler spectra and M-mode images from the MCA during different stages of the breath-hold maneuver. (A) Spectrum and M-mode at the beginning of breath-holding. Vertical yellow line in center denotes the start of breath-holding. (B) Spectrum and M-mode at the end of breath-holding. Vertical yellow line in center denotes the end of breath-holding when the subject inhales. (C) Spectrum and M-mode after the end of breath-holding, showing the decrease in flow velocity that persists for approximately 30 s after breath-holding. In all spectra, depth = 56 mm, gain = 50, sample volume = 8 mm, power = 420 mW/cm2, and filter = 100 Hz. Please click here to view a larger version of this figure.
Figure 7: Spectrum and M-mode from the MCA throughout breath-holding. Depth = 56 mm, gain = 50, sample volume = 8 mm, power = 420 mW/cm2, and filter = 100 Hz. Please click here to view a larger version of this figure.
Figure 8: Examples of bilateral spectra and M-mode images from the MCA. (A) Acceptable, but not optimal, spectrum and M-mode of the left MCA, with depth = 62 mm, gain = 69, sample volume = 12 mm, power = 420 mW/cm2, and filter = 100 Hz. (B) Good spectrum and M-mode of right MCA, with depth = 62 mm, gain = 56, sample volume = 12 mm, power = 420 mW/cm2, and filter = 100 Hz. (C) Good spectrum and M-mode of the left MCA. (D) Good spectrum and M-mode of the right MCA. For both (C) and (D), depth = 62 mm, gain = 56, sample volume = 12, power = 420 mW/cm2, and filter = 100 Hz. Please click here to view a larger version of this figure.
Age | Middle cerebral artery depth (mm) |
0–3 monthsa | 25 |
3–12 monthsa | 30 |
1–3 yearsa | 35–45 |
3–6 yearsa | 40–45 |
6–10 yearsa | 45–50 |
10–18 yearsa | 45–50 |
>18 yearsb | 50 |
Table 1: MCA depths at various ages. Sources: a = Bode25, b = Alexandrov et al.10
Critical steps in the protocol include 1) finding the MCA, 2) placing the headband, and 3) performing the breath-holding maneuver.
Modifications may be necessary depending on the subjects in the study. For example, subjects with Alzheimer’s disease may have difficulty following instructions, necessitating the use of a capnograph to ensure compliance with breath-holding instructions15. Young children may have difficulty following instructions and may be shy of the experimenter; hence, experimental protocols may need to be simplified for such a population (see Lohmann et al.2). Certain settings on the TCD machine may also need to be changed depending on the population of interest. For example, when insonating infants, who have thin cranial bones, reduce the power as much as possible, especially if TCD monitoring will take place over a period lasting several hours16.
Troubleshooting often centers around difficulties finding a good, stable TCD spectral signal. For example, for people older than 50 years of age, the temporal acoustic window becomes increasingly smaller as the age increases due to increased porosity of the bone of the cranium and tends to localize to the region just ahead of the ear (the “intertragal space”)12. In such a population, finding a good MCA spectral signal on both sides of the head may sometimes be impossible, and very slight changes in transducer angle or position may cause the signal to be lost. Because a good-quality signal is essential for experiments that depend on the envelope waveform for analysis, every effort should be made to increase MCA spectral signal intensity and quality. For instance, the gain can be adjusted to optimize the signal, and the sample volume can be increased to get a stronger signal. As a last resort, power may be increased. Finally, it is important to note that in approximately 10% of patients, the temporal acoustic window may be absent11,17. The temporal acoustic window can be readily found in infants and small children and is hardest to find in adults over the age of 50.
Limitations of fTCD include the acquisition of CBFV information at one spatial location17 rather than a wide field of view, albeit with very high temporal resolution. Thus, fTCD is a complement to fMRI, which gives cerebral hemodynamic information (and hence neural activity) with a wide field of view at a low temporal resolution18,19. Indeed, fTCD has a temporal resolution comparable to that of fNIRS20, with the important difference that fTCD measures hemodynamic changes at the level of the major cerebral arteries, whereas fNIRS measures changes in the cortex. Therefore, fTCD can fill in significant details about cerebral hemodynamic changes in response to neural activation, which no other neuroimaging modality is currently capable of measuring.
Potential applications of TCD include monitoring for cerebral embolus formation during cardiac surgery16 and monitoring to detect the outcome of tissue plasminogen activator treatment for stroke21. Potential applications of fTCD include any research question involving the neural response to internal or external stimuli, such as studying the lateralized processing of language in the human brain2,3,4, somatosensory “touch” stimulation5, or lateralization of visual processing6. In addition, fTCD can be used to study physiological (with or without neural activity changes) responses to stimuli such as exercise22 and breath-holding13,15,23. Finally, the low cost, portability, and simplicity of fTCD make imaging of large numbers of subjects practical, an advantage over fMRI and other neuroimaging modalities such as PET, e.g., when screening for preclinical Alzheimer’s disease15.
The authors have nothing to disclose.
This project is based on research that was partially supported by the Nebraska Agricultural Experiment Station with funding from the Hatch Act (Accession Number 0223605) through the USDA National Institute of Food and Agriculture.
Aquasonic | Parker Laboratories, Inc., Fairfield, NJ, USA | 01-50 | Ultrasound Gel |
Doppler Box X | DWL Compumedics Gmbh, Singen, Germany | Model "BoxX" | Transcranial Doppler with 2-MHz monitoring probes |
Kimwipes | Kimberly-Clark Professional | 34256 | Delicate Task Wipers |
Transeptic | Parker Laboratories, Inc., Fairfield, NJ, USA | 09-25 | Cleaning Spray |