Thanks for the details Kirk. Maybe the term "wall filter" isn't quite 
perfect since it can be separate from the DC blocker that must be 
present in any case to keep out the static "clutter". If either of these 
excludes much beyond 50 to 100 Hz as you suggest you will see more of a 
"dead zone" around 0 Hz where no flow information can get through in 
some late diastolic situations...


On 3/15/2016 3:00 PM, Kirk W Beach wrote:
> Dear Andrew,
> In pulsed ultrasound, from each transmit burst, you get a series of
> echoes from tissue;
> At 1 millisecond (ms) after transmit, the echoes from 0.75 mm return,
> at 10 ms echoes from 7.5 mm return.
>> From this series, you can make an A-mode (amplitude mode) display or
>> one line of a 2D B-mode (brightness mode) display.
> The Radio-Frequency (RF) echo series has a frequency range of 4
> MegaHertz to 6 MHz or maybe 2 MHz to 8 MHz.
> Each RF cycle at 5 MHz lasts 0.2 microseconds and represents 0.15 mm
> in depth.
> To get the B-mode, the amplitude of the echo series is measured at
> different depths, averaged over about 0.4 microseconds, providing a
> depth resolution of 0.3 mm.
> Higher ultrasound frequencies have better (smaller) depth resolution.
> The echoes from tissue are about 1000 times stronger (30 deciBels)
> than the echoes from blood, the specular echoes from arterial wall,
> when viewed at a 90 degree angle are about 10,000 times (40 dB)
> stronger than blood.
> They are like glare from sunlight off of a shiny surface.
> To perform Doppler, you need to repeat the pulse-echo sequence again,
> at least once, but usually the pulse-echo sequence is repeated 8 times
> (an ensemble of 8 for color Doppler) or 128 times (an ensemble of 128
> for spectral Doppler), all along the same line.
> In processing for Doppler, starting with the second pulse-echo
> sequence, the echoes from the previous pulse echo sequence are
> subtracted.
> So you get a series of (seven or 127) differences.
> Any solid tissue echoes like those from fat, muscle and wall, which
> are not moving, vanish in the subtraction, leaving only echoes from
> moving tissues in the difference signals such as  moving walls, and
> flowing blood.
> Only the echoes from moving walls and blood pass through this
> "Stationary Clutter" filter.
> But still, if there is wall motion, which will cause a Doppler shift,
> the wall signals pass through the clutter filter along with the moving
> blood signals.
> If the Doppler angle is not 90 degrees, then the wall echoes will not
> be bright specular, so they will be 30 dB above blood (rather than 40
> dB).
> If the Doppler angle is 0 degrees, the wall motion will be
> perpendicular to the ultrasound beam, and will provide no Doppler
> shift, and thus be blocked by the clutter filter.
> But, in normal examinations, the Doppler angle is between 30 degrees
> and 60 degrees, the Blood echo is weak but fast (with high frequency
> shift), the wall echo is slow (low frequency shift), but strong.  So,
> if you attenuate the low frequency signals delivered to your
> headphones, you will not hear the wall thump (low frequency) but will
> hear the blood flow Doppler signal, and if you attenuate the low
> frequency signals delivered to your spectrum analyzer, you will not
> see a bright band of low frequencies near the zero line on the
> spectral waveform.
> Generally, wall filters are set at 50 Hz or 100 Hz, because wall
> motion velocities are below 15 cm/s and blood velocities are generally
> higher.
> If the Doppler ensemble is 8 pulse-echo cycles, the wall filter can be
> designed to work pretty well, but it still might fail causing color
> "bleeding".
> Color bleeding can be suppressed by turning off the color in any image
> pixel that shows a strong echo (which must be wall).
> If the Doppler ensemble is 128 pulse-echo cycles, the wall filter can
> be designed to work better.
> BUT, if the case is INTERESTING, with turbulence and a bruit:
> Then the wall will vibrate with a high frequency (between 100 and 500
> Hz or more) and then the wall frequency will pass through the filter
> and be superimposed on the spectral waveform.  It will differ from the
> blood flow signal: the vibration signal is double sideband, showing up
> symetrically above and below the ZERO baseline, compared to the blood
> flow signal which is single sideband, showing up only above or below
> the ZERO line.
> The spectral display is usually adjusted in brightness so that the
> brightest signal shows as "white" and signals within 20 dB of the
> brightest are shown as gray.
> So, if the wall filter is off, then the wall motion will be used to
> define "white" and the weaker blood signal (25 dB weaker = dark gray)
> might be hard to see.
> If the wall filter is on, then the wall motion will be absent in the
> display and the blood signal will be used to define "white" making the
> spectral waveform easily visible.
> Kirk
> On Mon, 14 Mar 2016, Andrew Bebry wrote:
>> My student is questioning the following:
>> Which signals are eliminated by the wall filter in a spectral display?
>> Andy Bebry
>> To unsubscribe or search other topics on UVM Flownet link to:
> To unsubscribe or search other topics on UVM Flownet link to:

Buster Greene and Associates, LLC

To unsubscribe or search other topics on UVM Flownet link to: