Estimating Pitch Information From Simulated Cochlear Implant Signals With Deep Neural Networks.
Takanori Ashihara, Shigeto Furukawa, Makio Kashino
PMID 39569552WHAT IT FOUND
Simulated cochlear implant signals contain enough pitch information for accurate estimation if at least eight channels are used.
Background noise significantly degrades this accuracy, particularly when pulse rates are low. This suggests device settings, not just patient factors, limit pitch perception in noisy environments.
Key findings
01In quiet conditions, simulated cochlear implant signals with eight or more channels allowed the model to estimate fundamental frequency (pitch) with accuracy approaching that of the raw audio signal.
02Background noise significantly reduced pitch estimation accuracy for all models, with lower pulse rates causing larger errors than higher pulse rates.
03To achieve a 75% accuracy rate for pitch estimation in noisy conditions, a configuration with 600 pulses per second required only four channels, whereas 400 pulses per second required twelve channels.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used a computational model (DNN), not human listeners. The model's ability to extract information does not guarantee that a human auditory system can perceive it. The simulation used the CIS strategy and did not test other modern strategies like fine-structure processing. The noise used was stationary environmental noise (e.g., factory, car) and did not include competing speech or harmonic noise, which might affect pitch perception differently. The model did not account for biological factors such as neural survival, electrode insertion depth, or auditory plasticity.
Declared interests
The authors declared no conflicts of interest. The study was supported by non-U.S. government funding.
The easy way to misread this
Do not assume that because the computational model accurately estimated pitch, human CI users will perceive pitch with the same fidelity. The model lacks the biological constraints and neural limitations of actual patients.