Cochlear Implant Processing Strategies

Question 1

What is a processing strategy?

Answer 1

• Set of rules and processes for converting the acoustic input signal into electrical stimulating waveforms that are applied as the output by a cochlear prosthesis
• Can be paired stimulation or sequential
• Captures the distinguishing parameters of sound

Question 2

What are the distinguishing patterns of sound?

Answer 2

• Temporal domain
• Intensity domain
• Spectral domain

Question 3

What is the intensity domain?

Answer 3

-Influenced by how sound is captured and how it is processed by the implant system

Influenced by:

• Directional microphones
• Mic placement
• AGC
• IDR

Question 4

What is the purpose of AGC?

Answer 4

• Boost soft sounds

- Compress loud sounds

Question 5

What is the spectral domain?

Answer 5

• NH: ~20 to 20,0000 Hz filtered by the BM
• Implant system: filters are assigned to locations across the electrode array
• CIs process ~250 to 8000 Hz
• Not as much resolution is needed for speech understanding (vowels can still be distinguished even with “sloppy” formants)
• Music is different because the targets are points, not ellipses, in frequency space (notes that are close in frequency can be processed through the same CI filter bank = perceived as the same)

Question 6

What is Continuous Interleaved Sampling (CIS)?

Answer 6

• One of the first/most simple strategies
• May be helpful for patients with cognitive issues
• Filters sounds using BP filters
• Envelope variations at different bands are represented at corresponding electrodes by modulating biphasic pulse trains
• Envelope is compressed nonlinearly to maintain the 100 dB DR of sounds
• Pulses are sent sequentially to electrodes in the appropriate frequency place of the cochlea (~1000 pps)
• Output of LP filters is sent to a compressor that uses a log function to map the input signal into the EDR
• Output of the compressor is sent to a pulse generator with a fixed frequency

Question 7

What are available processing strategies?

Answer 7

• Continuous Interleaved Sampling (CIS)
• Spectral Peak (SPEAK)
• High Resolution (HiRes)
• Advanced Combination Encoder (ACE)
• Fine Structure Processing (FSP)

Question 8

What is HiRes?

Answer 8

• AB
• Uses a high stim rate (up to 5000 pps) and up to 16 channels (double the 8 of CIS)
• Does not use the LP filter for envelope detection, but a half-wave rectifier
• Uses current steering, varying stimulation on multiple electrodes in different patterns for different frequencies
• More sophisticated AGC algorithm

Question 9

What is Advanced Combination Encoder (ACE)?

Answer 9

• Based on “n of m” processing strategies
• Channel selection scheme
• Envelope signals for the different channels are scanned prior to each frame of stimulation
• The highest amplitudes are identified and stimulus is delivered to those electrodes
• The parameter is fixed for ACE and variable for SPEAK
• Stim rate: ACE = 1000 pps; SPEAK = 250 pps

Question 10

What are “n of m” processing strategies?

Answer 10

• Acoustic energy in each of m channels is determined and stim is administered to only the n channels with the highest amplitude inputs
• N = maxima
• Reductions of the number of active electrodes results in: 1) faster stim rates, 2) reduction of channel interactions and masking, and 3) increased battery life
• Occasionally, desired info lies in channels with lower amplitudes and is therefore not transmitted

Question 11

What is Fine Structure Processing (FSP)?

Answer 11

• MED-EL
• Novel form of CIS
• Intermediate pitches generated via bell-shaped, overlapping filters to provide better spectral resolution
• Designed to improve perception of HF phonemes which are identified based on place cues (/s, t, f/)
• Modulated timing of channels specific sampling in the LF region depending on input freqs in that region
• Theoretically, improves speech recognition, sound quality, recognition of vocal pitch, music appreciation, and music recognition