RFcap: A software a
Contents
1. Biomed Eng 1 115 142 14 231 24 25 2d2 283 29 Figure captions Figure 1 A typical system configuration for capturing and analyzing RF signals An acoustic signal is processed by the speech processor A and coded as an RF signal which instead of being sent to a patient s implant is sent to the receiver coil of the PCI B The hardware continuously decodes the incoming RF signal storing the data in hardware buffers which are read by RFcap software on a computer C connected to the PCI via a Nucleus IF6 card Figure 2 The captured and decoded stimulus activity for the utterance asa is displayed onscreen as an electrodogram Note that the centre of gravity CGrav of the electrodogram is computed as the modal electrode and its corresponding modal stimulation level Figures 3 to 9 are also based on this same signal Figure 3 Electrode histograms show the distribution of activity across the array of 22 intracochlear electrodes for the entire signal The lower histogram is weighted by the stimulus intensity computed as a percentage of the dynamic range between T and C levels for that electrode Figure 4 The activity on a selected electrode or channel is summarized here showing the distribution of stimulus intensities in Current Level units as well as the instantaneous stimulation rate between every 2 consecutive stimuli on the same electrode The x axis of the upper graph is in Current Level units while th2. Period us Powerup Periods Pulse by pulse n 75 on 0 1000 Number of occurences sa 600 wo 200 0 500 1000 1500 2000 2500 3000 3500 Pulse Number Distribution of Powerup Periods us on 0 Median Prd 399us 2506 Hz 30 20 0 100 200 300 400 500 600 700 800 900 1000 Powerup Period us Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 8 RFcap R PICNIC RFcap Captured ACE _MP3000 asa_spla_mp3000_map190 seq hee ce clea moat pi Won bia a Mu AIA ALi i ut iW iva tiaw eae tina a i etna aalt ddillad a Pea torn ANILERE T E T Wt see ee ee anat sacs tii tind Bitia tah ona eC WD Me ae amen 1 HU Ma a mm n PT ddl d n E O OE O tae i ow ddd Teel ae er er eee Il A T T le ETA OCOLIT il wil Aut MRAN mo a aa a oa A A ieee DME ti a a PAET hA A A b h tun na Tee eee me aaa Rae y Se Wibin ere A Modan poo 14 bud hha bein we a MD Idee ahili in i Man WW Matin i wll MVE moi ee mo EE E S rT nMi hi bonn LIH peee thaws tide on i elu a im ine Vit Hill team a _ het oe biun Wak u I Mah oe bwli Sain d5 pa Playback sound and capt Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contac
3. decode functionality to provide a practical means for displaying the CI stimuli together with a number of analysis and diagnostic functions Although NIC and NMT are only available to collaborative research groups working with the Nucleus CI24M R CI24RE and CI500 cochlear implants the general principle of signal capture analysis and diagnostics is applicable to other cochlear implant systems 231 24 25 2d2 283 29 System Description Typically a CI speech processor receives its input signals via its microphone or from a direct connection to a separate signal source The input signal is then processed according to a particular coding strategy resulting in a sequence of stimuli which is then encoded into RF signals for transmission to the target CI Instead of being sent to the implant of a CI user the RF signals can be intercepted by specific hardware capable of receiving and decoding these signals The decoded information can then be transferred to software for subsequent analysis as well as storage Figure 1 illustrates a typical configuration of such a capture and analysis system lt lt Figure 1 about here gt gt Suitable hardware for the Nucleus CI24M R CI24RE and CI500 cochlear implants consists of the PCI unit together with the IF6 interface card both of which belong to standard clinical hardware although no longer the latest for the Nucleus cochlear implant system The PCI contains the necessary circuitry for receiving
4. field or in a soundproof test box It is unusual to use the direct accessory input as this bypasses the microphone input However direct input is useful for testing with signals such as pure tone sinusoids when an anechoic test box is not available c Using the Frequency Response function the user can quickly check if the frequency response of the speech processor is according to specifications The discrete nature of the mapping into Current Level units limits the output resolution somewhat but this function will nevertheless be able to detect large deviations from the expected norm Note that the PCI IF6 hardware is not compatible with older Nucleus 22 implants Diagnostics for Nucleus 22 implants will have to be made using either sCILab or a digital oscilloscope It should also be kept in mind that it is not possible to check the output of Nucleus Freedom or Nucleus CP810 sound processors programmed for CI24RE or CI500 implants without reprogramming the processors with a map for the CI24M R implant when testing in conjunction with the PCI IF6 hardware 12 Summary In summary RFcap is a software analysis tool which is capable of capturing and storing the decoded output RF signals from Nucleus speech processors for the CI24M R CI24RE and CI500 cochlear implants Such an analysis tool can be applied to many situations ranging from basic research developments to clinical diagnostics 13 231 24 25 2d2 283 29 Reference
5. please contact the Editor for cost details Manuscript_110214_Fig Figure 3 Electrode Activity loj x File Edit View Insert Tools Desktop Window Help Deb B QAQMea 2 O0H ab Electrode Activity Unweighted 300 200 100 Number of occurences 22 21 20 19 18 17 1615 1413121110 9 8 7 6 43 2 1 Electrode Electrode Activity Weighted by DR Map 84 test patient 24 Right 300 200 100 Weighted number of occurences 22 21 20 19 18 17 16 15 1413121110 9 8 7 6 4 3 2 1 Electrode Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 4 l gt Channel Activity Diagnostics ST File x Dee SlQanea e oe Distribution of CLs nPulses 205 CGray 106 64 CL 80 to 120 Je20 50 40 30 20 10 80 85 90 95 100 105 110 115 120 Channel Rate Hz nPulses 205 Rate Mode 502 Hz 9 to 558 ooogoo0o0 oo o D SO O OD OOOO O O oc0oo co O oO o o o 0 500 1000 1500 2000 2500 3000 3500 Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 5 Stimulation Rate Diagnostics aink File Edit View Insert Tools Desktop Window Help x Dee Sls QAMo
6. segments The modal electrode number with the largest intensity within each segment and the modal stimulation level in Current Level units within each segment are then summarized for each segment Here the results from 5 sweeps at different intensity levels are shown combined in the same display 16 Non colour figure Click here to download Non colour figure Manuscript_110214_Figure_01 doc Figure Host PC with Matlab NIC NMT amp RFcap software to IF6 Card 6 PCI ree OWE Speech Acoustic Processor Signal Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 2 RF cap R PICNIC RFcap Documentation Paper freedomSP_map84_ ACE asal dseg 4 ME tim l CE CILORR OTETI I EE E da lidhin dildan hi a ma wia brb hal A ba j Me d UN tL Ar Pe EL Aia ma TEOR EE E PRIMM GOT ADunn dka nkannd wundan a A Viv tomate ee MO TEM MT Le si rte ests eae dee A eT er d H b LIII i TT prs onani I Is 1 om Heidi abate tseama of toda 1 Woa C E T Ce E A EE ee lf Att bith ow LETIGUT SEG PE r Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies
7. E 08 eo0 Stimulation Rates pulse by pulse 6000 4000 2000 Rate Hz per pulse 0 500 1000 1500 2000 2500 3000 3500 Pulse Number Stimulation Rates Per set of maximas 1000 500 Rate Hz per MaxSet 0 50 100 150 200 250 300 350 400 Maxima Set Number Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure Maximas Diagnostics Edit View File 6 Insert Tools Desktop Window Help Oe Ss Qarya e 08 seo Count period us NumMax per Set 2000 wn o Oo 1000 on oO Maximas Counted basal apical order 100 200 Maxima Set Duration us of count periods 300 400 oO 100 200 Maxima Set 300 400 Number of Occurences Number of occurences O x Mu Distribution of Maximas 400 300 200 100 0 0 2 4 6 8 10 NumMax per Set Distribution of count periods us 400 300 200 100 0 500 1000 1500 2000 Count Period us Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 7 F rowerup Diagnostics o Mu File Edit View Insert Tools Desktop Window Help Dea Ss RAarmo e 08 aeo Powerup
8. Editorial Manager tm for Cochlear Implants International Manuscript Draft Manuscript Number CIM58R2 Title RFcap A software analysis tool for multichannel cochlear implant signals Article Type Technical Note Keywords Cochlear Implant Electrodogram Nucleus Implant Communicator Nucleus Matlab Toolbox sCILab Corresponding Author Wai Kong Lai PhD Corresponding Author s Institution University Hospital Zurich First Author Wai Kong Lai PhD Order of Authors Wai Kong Lai PhD Norbert Dillier Manuscript Region of Origin SWITZERLAND Abstract Being able to display and analyze the output of a speech processor which encodes the parameters of complex stimuli to be presented by a cochlear implant CI is useful for software and hardware development as well as for diagnostic purposes This firstly requires appropriate hardware which is able to receive and decode the radio frequency RF coded signals and then processing the decoded data using suitable software The PCI IF6 clinical hardware for the Nucleus CI system together with the Nucleus Implant Communicator and Nucleus Matlab Toolbox research software libraries provide the necessary functionality RFcap is a standalone Matlab application that encapsulates the relevant functions to capture display anda nalyze the RF coded signals intended for the Nucleus CI24M R CI24RE and CI500 multichannel cochlear implants Manuscript DO NOT INCLUDE ANY AUTHOR NAMES OR CONTACT DETAILS IN THIS FI
9. LE Click here to download Manuscript DO NOT INCLUDE ANY AUTHOR NAMES OR CONTACT DETAILS IN THIS FILE Manuscript_1 1 Introduction 12 The ability to display and analyze the output of the speech processor which encodes the stimuli to 43 be delivered by a cochlear implant CI can be extremely helpful and even essential in many 64 instances both in research as well as clinical settings For example when developing new speech 95 coding strategies the output of the new algorithms need to be examined systematically and 116 thoroughly This is not difficult with simple inputs but it becomes cumbersome when complex 147 inputs such as real speech tokens are used Displaying the complex signal output makes it possible 168 to examine this output for patterns that may otherwise not be easily observed when using simple input signals Another example in a clinical environment is to examine the output of a speech 210 processor for diagnosing hardware problems such as defective microphones or accessories a With the present generation of commercial cochlear implant systems approved for clinical use 242 stimulus parameters are transmitted via radio frequency RF coded signals transcutaneously to the Sd3 cochlear implant By capturing and decoding these RF coded signals it is possible to examine the 3114 output of the entire signal processing chain 335 Capturing these RF coded signals requires appropriate custom hardware as well as software 5116 Previously th
10. a CI24M R implant This restriction does not apply to the SPrint ESPrit and ESPrit3G speech processors which do not check for the implant type Software Requirements Although RFcap is a Matlab application it has also been compiled as a standalone application which can be executed on a PC which does not have Matlab installed on it RFcap was developed using Matlab version r2007b and has also been successfully tested with version r2009b Since RFcap also permits the user to connect to the Nucleus Custom Sound clinical patient database the Database Toolbox from Matlab is also required for the database connection to function properly The functionality to communicate with the capture and decoding hardware is provided by the NIC and NMT research software libraries from Cochlear and the latest versions presently v2 23 and v4 31 respectively are required Use of the NIC and NMT libraries are subject to agreement with Cochlear 231 24 25 2d2 283 29 RF cap Functions Once RFcap is running the PCI continuously decodes all RF input signals and stores the corresponding data in the hardware buffers as frames The data transfer from the hardware is controlled by RFcap and is specified either in terms of the desired number of frames or as a time interval RFcap initiates the transfer and this is stopped when either the specified number of frames has been transferred or the specified time interval has elapsed whichever occurs first The maximu
11. and decoding the RF signals The decoded information consists of a series of parameters describing each biphasic pulse in the stimulus sequence being sent to the target cochlear implant Each set of parameters i e active and reference electrode phase width interphase gap and duration till the next stimulus defining a single stimulus pulse make up a frame The decoded frames are continuously stored and refreshed in hardware data buffers whose contents can in turn be read using the appropriate software functions provided by the NIC and NMT libraries RFcap is such a software application which uses these functions to transfer the stimulus data from the hardware buffers into the PC Once transferred RFcap can process the data for visual display as well as for analysis Hardware Limitations More recent speech processor models equipped with telemetry capability e g the Nucleus Freedom or the Nucleus CP810 Sound Processors may check the implant type to which it is transmitting and automatically stop its RF transmission if the expected implant type is not found The PCI which is equipped with a CI24M R implant chip would make it impossible to check the output of Freedom or CP810 processors programmed for CI24RE or C1500 implants This behaviour is defined in the speech processor s firmware and as such can be circumvented either by using a version of the firmware that does not behave this way or by reloading the processor with a map for
12. ayed superimposed over one another for comparison The second stimulus data set is loaded from file as a Reference Signal and the two electrodograms are shown in different colours The electrodograms in Figure 8 are for the sound token asa obtained with the MP3000 coding strategy Noguiera et al 2005 in the foreground dark blue and the ACE coding strategy in the background light blue It can be seen that MP3000 produces less activity but still has a very similar spectral representation The horizontal slider allows the user to shift the shorter of the two electrodograms left or right to align the patterns when two stimulus data sets are being displayed lt lt Figure 8 about here gt gt When a Reference Signal is loaded the Electrode Activity as well as Channel Activity analysis functions also display the results from the Reference Signal superimposed in the background in the same display Figure 9 show the corresponding electrode activity for the electrodograms from 4d8 Figure 8 confirming that the MP3000 activity pink or red is less in number as well as intensity than the ACE activity light blue lt lt Figure 9 about here gt gt RFcap also provides a diagnostic function to check the frequency response of a speech processor Basically this involves presenting a discrete frequency sweep consisting of 7 consecutive pure tone bursts 125 250 500 1000 2000 4000 and 6000Hz respectively of equal lengt
13. e 20 were well distributed across the dynamic range between the T and C levels and the 4d8 instantaneous channel stimulation rate corresponded well to the expected 500Hz 502Hz was computed due to discretization from the map used to generate the stimulus sequence Fluctuations around this value are due to the order in which the sets of maxima were selected When a particular channel is not repeatedly selected the time interval to the next occurrence becomes larger resulting in a lower instantaneous rate lt lt Figure 4 about here gt gt The instantaneous pulse by pulse stimulation rate can be computed from the time interval between any two consecutive stimulus pulses including powerup pulses in the stimulus sequence Generally this should match the overall stimulation rate for the corresponding map given by the product of the channel stimulation rate and the number of maxima and is expected to be constant If the input signal is too weak to yield the specified number of maxima the signal is padded with powerup pulses The ACE coding strategy selects the specified number of maxima from the input signal and presents them in a basal to apical order at the channel stimulation rate Thus the channel stimulation rate can also be determined by examining the stimulus sequence for sets of maxima and computing the rate at which these maxima sets are presented Powerup pulses are ignored when determining maxima sets Weak input signals resul
14. e x axis of the lower graph is the frame number of the first stimulus in the pair used to compute the instantaneous channel stimulation rate Figure 5 The overall instantaneous stimulation rate is computed from the time interval between each consecutive stimulus i e including powerup pulses regardless of the corresponding electrode The stimulation rate per set of maxima assumes that the selected maxima within a set are presented in a basal to apical order This should correspond to the channel stimulation rate 15 4d8 Figure 6 The number of maxima found in each set and the duration of each set are summarized here as well as their respective distributions Figure 7 The occurrences of powerup pulses and the distribution of their corresponding durations are summarized here Figure 8 Two superimposed electrodograms in different colours allow for visual comparisons In this example the activity in the foreground dark blue from the MP3000 coding strategy can be seen to be less than the activity in the background light blue from the ACE coding strategy but still has a very similar spectral representation Figure 9 The electrode histograms for the signals in Figure 9 confirm that MP3000 has less activity number of stimuli and this has generally less intensity weighted number of stimuli than the corresponding ACE activity Figure 10 The frequency response assumes that the input signal consists of 7 equal time
15. hs e g 500ms and intensity to a speech processor and then capturing the corresponding output activity The sound playback level is assumed to be appropriately calibrated and as pure tones are being used the playback should be in anechoic surroundings for instance a soundproof test box The captured stimulus data are divided into seven corresponding equal time segments and for each segment the modal electrode number assuming that this is the central electrode if the activity is spread over several electrodes as well as the modal stimulus intensity for that electrode is computed and summarized onscreen Note that the frequency response here refers to the input output function for the entire signal path including the microphone plus any additional filtering as well as threshold clipping and discretization associated when mapping the input signal into Current Level units Measurements at different intensity levels e g varying in 5dB steps can also be combined into a single diagnostic display as shown in Figure 10 lt lt Figure 10 about here gt gt Ideally when measuring the frequency response an artificial map with all electrodes having the same T and C levels should be loaded into the speech processor as well as RFcap 10 231 24 25 2d2 283 29 RF cap as a development and research tool RFcap allows for easy comparisons of new speech coding strategies against existing ones For instance the MP3000 coding stra
16. ight blue to dark blue graduated scale The colours can be manually altered by the user to optimize the visual appearance of the electrodogram Some of the time information in the stimulus sequence is simplified in the electrodogram display For instance 4d8 biphasic pulses with finite phase widths and interphase gaps are represented simply by vertical bars with a fixed onscreen width of 1 pixel Thus timing information such as the phase width and interphase gap although available in the decoded data are not represented visually in the electrodogram by the width of the vertical bar The time interval between stimuli however is correctly displayed A summary of the data s stimulus parameters is also provided below the electrodogram Powerup pulses are represented as stimuli presented on electrode 0 lt lt Figure 2 about here gt gt The height of each vertical bar is scaled as a percentage between the minimum and maximum Current Level values possible for each electrode for instance as defined by the Threshold and Comfortable levels T and C levels of a speech processor map RFcap allows the user to connect to the clinical Custom Sound database to read these speech processor maps When a map is loaded RFcap will use the map s corresponding T and C levels to encode the height of the vertical bars If a map is not loaded RFcap simply assumes the T and C levels to be 0 and 255 respectively It is up to the user to load the correct ma
17. is was possible for the Nucleus family of cochlear implants using sCILab Lai et al 387 2003 a 16 bit Borland Pascal application which communicated with standard clinical hardware in 78 particular with the Nucleus DPI IF4 Dual Processor Interface and Nucleus IF4 interface card and 149 PCI IF5 Processor Control Interface and Nucleus IF5 interface card hardware combinations 720 However these clinical hardware combinations have since been superseded and the subsequent 12 hardware combination of PCI with the Nucleus IF6 interface card is not compatible with sCILab 5Q2 Furthermore the latest clinical hardware with a USB programming Pod does not have any decoding 233 or storage capability 524 The PCI IF6 combination while no longer the latest clinical hardware for the Nucleus CI system 725 represents the most recent and still available hardware configuration able to continuously decode 59 626 and store the RF signals The software functionality to communicate with the hardware and thus capture the decoded data is provided by the Nucleus Implant Communicator NIC and the Nucleus Matlab Toolbox NMT research software libraries Irwin 2006 Swanson and Mauch 2006 from Cochlear NIC comprises a C C library which is called by NMT the latter of which is intended for the Matlab by Mathworks software environment RFcap is a standalone Matlab application designed and written by the first author which similar to sCILab utilizes this capture and
18. m number of frames that can be transferred in a single capture is 65535 frames The corresponding time interval depends on the duration of each frame which is in turn defined by the overall stimulation rate of the speech processor map used to generate the stimulus sequence e g a 500Hz ACE map with 10 maxima overall rate of 5000Hz has a frame period of 200us RFcap terminates the data transfer if no decoded RF signal is available for a specified time interval timeout RFcap can also control the playback of an arbitrary sound token over the PC s sound system Since the data transfer is controlled by RFcap synchronized playback of sound tokens and capture of the resultant RF signals is also possible By default RFcap sets the time interval to the length of the sound token The user can still adjust the number of frames to be captured as desired The transferred data is displayed in the form of an electrodogram which very closely resembles a spectrogram Figure 2 illustrates such an electrodogram for the input utterance asa as captured from a Freedom SP using a 500Hz ACE map with 10 maxima The elapsed time is shown on the x axis and the electrode number in increasing order of its frequency mapping on the y axis Each stimulus is shown as a vertical bar whose height is proportional to the corresponding stimulus intensity The height of the vertical bar is also colour coded using a 10 step two colour the example in Figure 2 ranges from l
19. mply counting the number of stimuli presented on a particular electrode or each stimulus on that electrode can be weighted according to its intensity before being added together The weighted histogram gives a better impression of the relative signal intensity on each of the channels The weights are linearly distributed between 0 0 and 1 0 Ifa map is loaded weighting will be relative to the T and C levels of the corresponding electrode A stimulus intensity at T level is given a weighting of 0 0 while a stimulus intensity at C level corresponds to a weighting of 1 0 If a map is not loaded weighting will be relative to the full range of 0 to 255 Current Levels For instance a stimulus intensity of 153 Current Level units would yield a weight of 153 255 0 6 The upper and lower electrode histograms in Figure 3 respectively show that in general the distribution of unweighted and weighted activity are similar Closer examination shows that although more stimuli were presented on electrode 7 than electrode 5 the stimuli on electrode 7 were generally weaker thus yielding a smaller weighted sum than that on electrode 5 lt lt Figure 3 about here gt gt Additionally the distribution of the stimulus intensities in Current Level units on a selected electrode or channel can be examined as well as the instantaneous channel stimulation rate between two consecutive stimuli on that electrode Figure 4 shows that the stimuli presented on electrod
20. p into RFcap which corresponds to the one being used in the speech processor at the time the RF capture is being performed Loading the wrong map into RFcap will result in erroneous decoding of the signal intensities for instance captured stimulus intensity values may lie outside the range given by the T and C levels in the loaded map The loaded map also contains other parameters such as the phase width interphase gap the stimulation rate and the number of maxima which can also be used to cross check the captured and decoded RF stimulus data for mismatches or errors RFcap provides such a basic check of the captured RF data and reports if there are mismatches in these parameters By default RFcap displays the entire captured signal scaling the horizontal time axis accordingly However the user can also examine the electrodogram in detail by zooming in and out on different portions of the electrodogram 4d8 Once the stimulus data are available in digital form they can be saved to file Saved data can be read into RFcap for analysis in the same manner as data that have just been captured Analyzing the stimulus data The stimulus data can be analyzed in various ways as described below The corresponding examples are all based on the captured stimulus data for the utterance asa as shown in Figure 2 The amount of activity on each of 22 electrodes over the entire signal is summarized in the form of electrode histograms either by si
21. s Henry B A and Turner C W 2003 The resolution of complex spectral patterns by cochlear implant and normal hearing listeners J Acoust Soc Am 113 5 2861 73 Irwin C 2006 NIC v2 Software Interface Specification E11318RD Technical Report Lane Cove NSW Australia Cochlear Ltd Lai W K B gli H and Dillier N 2003 A software tool for analyzing multichannel cochlear implant signals Ear and Hearing Baltimore MD 24 5 380 391 Miiller Deile J 2009 Speech intelligibility tests in cochlear implant patients HNO 57 6 580 92 Noguiera W B chner A Lenarz T and Edler B 2005 A psychoacoustic NofM type speech coding strategy for cochlear implants EURASIP Journal of Applied Signal Processing 18 3044 3059 Sagi E Kaiser A R Meyer T A and Svirsky M A 2009 The effect of temporal gap identification on speech perception by users of cochlear implants Journal of Speech Language amp Hearing Research 52 2 385 95 Sagi E Meyer T A Kaiser A R Teoh S W and Svirsky M A 2010 A mathematical model of vowel identification by users of cochlear implants J Acoust Soc Am 127 2 1069 83 Swanson B A and Mauch H 2006 Nucleus Matlab Toolbox 4 20 Software User Manual Lane Cove NSW Australia Cochlear Ltd Zeng F G Rebscher S Harrison W V Sun X and Feng H 2008 Cochlear Implants System Design Integration and Evaluation IEEE Rev
22. such as the Nucleus USB Programming Pod the PCI IF6 hardware is all the more valuable as it represents the last widely available clinical equipment with full decoding and buffering capability which RFcap is able to take advantage of 11 Ai 24 25 2d2 27 293 29 30 3114 345 34 3 36 387 39 4 118 42 439 44 490 46 47 agl 502 59 626 RF cap as a diagnostic tool RFcap can also be used in the clinical environment to diagnose problems with speech processors and their related accessories Occasionally CI patients have complaints of background noises in the speech processor or that the sound has become muffled With RFcap it is possible to quickly test the output of the speech processor before sending the unit for repair There is at present no such diagnostic tool available for testing in a CI clinic for the current generation of Nucleus speech processors RF cap allows the following diagnostics to be performed a By capturing the RF activity without any signal input spontaneous noise can be detected and documented For instance when the telecoil of a speech processor is enabled it has been reported that this can produce spurious background noise which may irritate the CI user b Testing the output with a given input signal particularly one calibrated at a known level allows the tester to compare the resultant activity with expected electrodograms The sound source level is typically calibrated be it free
23. t the Editor for cost details Manuscript_110214_Fig Figure 9 Figure 1 Electrode Activity 10 Fie Edit view Insert Tools Desktop Window Help De eS s QAavmoa e 08 so0 RAPICNIC RF cap Captured ACE_MP3000 asa_sp1a_mp3000_map190 seq Electrode Activity Unweighted 400 300 200 100 Number of occurences 22 21 20 19 18 17 1615 14131211109 8 7 6 5 43 2 1 Electrode Electrode Activity Weighted by DR Map 190 test freedom Left 250 200 150 100 50 Weighted number of occurences 22 21 2019 18 17 1615 14131211109 8 7 6 5 43 2 1 Electrode Colour figure print author charge applies please contact the Editor for cost details Click here to download Colour figure print author charge applies please contact the Editor for cost details Manuscript_110214_Fig Figure 10 Figure 1 Freq Response Diagnostics Me x File Edit View Insert Tools Desktop Window Help DSak aana enal g Discrete Sweep Frequency Response 250 200 150 Current Level 100 50 125 250 500 1000 2000 4000 6000 Frequency Hz
24. tegy based on forward masking effects Noguiera et al 2005 results in a reduction in the overall amount of stimulus activity but should retain the overall spectral representation Figures 8 and 9 show how the MP3000 output compares to the standard ACE coding strategy output The electrodograms have the same spectral extent but the MP3000 signal is clearly more sparse Although it is also possible to examine the decoded signals using implant in a box hardware together with a multichannel digital oscilloscope the analysis options and the maximum number of channels are often limited RFcap has the advantage that it is able to capture and display the activity on all 22 stimulation channels using hardware that is already available in the CI clinic Custom analysis functions for the captured data can also be created as required whereas the data analysis functions available with a digital oscilloscope may not necessarily be suitable The predecessor sCILab software continues to be used and cited in various cochlear implant studies e g Henry and Turner 2003 Miiller Deile 2009 Sagi et al 2009 Sagi et al 2010 Zeng et al 2008 Other examples are cited in Lai et al 2003 With the old clinical hardware now already superseded the need for a similar software that works with more recent clinical hardware is quite urgent and RFcap is intended to fulfil this need As the present trend continues towards even smaller and simpler interface hardware
25. ting in less than the nominal number of maxima specified in the corresponding speech processor map will result in shorter time intervals for that maxima set and accordingly a higher channel stimulation rate Irregularities in the instantaneous stimulation rate and the channel stimulation rate can be seen in the onset and offset portions of the utterance asa as shown in Figure 5 lt lt Figure 5 about here gt gt The maxima sets found in a stimulus sequence can be summarized according to the number of maxima in each set as well as their corresponding duration as shown in Figure 6 4d8 lt lt Figure 6 about here gt gt When there is no signal input and hence no signal output to the implant the speech processor nevertheless continues to send RF pulses to the implant in order to keep the implant powered up These powerup pulses produce no stimulation with a stimulation level of OCL However these powerup pulses are detectable in the RF signal stream The time interval period between two consecutive powerup pulses as well as the distribution of these powerup periods for the utterance asa are shown in Figure 7 The various powerup periods found in the signal are primarily responsible for the fluctuations in the instantaneous stimulation rate and channel stimulation rate observed in Figure 5 lt lt Figure 7 about here gt gt RF cap allows the electrodograms from two different sets of stimulus data to be displ
Download Pdf Manuals
Related Search