| Home | About Us | Contact Us |
|
Envirograms and Artificial IntelligenceDouglas L. Beck Au.D. Donald J. Schum, Ph.D./CCC-A A "soundscape" is an acoustic environment. Soundscapes are the sounds one hears daily as they traverse work, recreation and home experiences. Each person's soundscape is unique and indeed, most people experience a multitude of soundscapes across their daily activities. Because soundscapes are dynamic, effective amplification must also be dynamic. Patients with sensorineural hearing loss (SNHL) have unique, personal auditory perceptions. Their perception of sound is often distorted, secondary to; their type and degree of hearing loss and is subject to variation in pure-tone and speech thresholds, comfortable and maximum listening levels, auditory processing ability, cognitive abilities, history, expectations, needs, emotional status, and importantly -- the "soundscapes" they experience. In essence, each individual experiences the world of sound within their own limits and abilities across unique and highly variable acoustic environments. As professionals, we have typically characterized our patient's hearing ability and hearing loss. However, we have not typically quantified the patients' individual and variable acoustic environments to better understand their specific auditory demands, or to recommend specific amplification strategies based on their unique listening needs. In other words, until recently, we have not had the ability to quantify individual soundscapes. The Activity Analyzer (AA) and the Sound Activity Meter (SAM) are new advanced data-logging devices which go beyond "traditional" hearing aid-specific data-logging characteristics. The AA and SAM record acoustic environments in terms of sound pressure levels with respect to quiet, noise, speech and speech-in-noise. As a result of these recordings, Envirograms can be created. In this article, we'll discuss how Envirograms can be used to demonstrate how advanced automatic decision making, made possible through Artificial Intelligence (AI) in advanced technology hearing aids, addresses the specific communication needs of the patient, based on their personal Envirogram. I. Creating the Envirogram: The Activity Analyzer (AA) and the Sound Activity Meter (SAM). The Activity Analyzer (AA) was released in the Spring of 2005 within the Oticon Syncro 2. The AA records traditional data-logging information such as wear time, program selection and volume control usage. However, advanced analysis of the acoustic environment based on quiet, speech, noise and speech-in-noise, as well as the directional mode of operation, such as; surround/omni, split or full directionality is also acquired. For example, the AA in a Syncro-based recording might indicate the instrument was in full, split or surround/omni mode, 20, 35 and 45 percent of the time, respectively.
More recently (Fall, 2005) we introduced a self-contained, tie-tack sized version of the AA, referred to as the Sound Activity Meter (SAM). The SAM (Figure 1) records a complete, accurate and objective sound analysis of the soundscapes experienced by the patient. Quiet, noise, speech and speech-in-noise conditions are captured in terms of percentages-of-time. The SAM is designed to be worn by a prospective patient before a hearing aid is fit to help motivate the patient as to the need for advanced signal processing and to provide important input to the fitting and orientation process. Note: SAM and AA do not record actual conversations, they record only sound pressure levels and the type of sound environment. Both tools (AA and SAM) can be accessed and their reports printed through our e-CAPS and Genie software systems. II. The Envirogram: The information recorded via the AA or the SAM is displayed on a graph called the "Envirogram." The vertical axis of the Envirogram represents "percentage of total time." The horizontal axis reflects sound pressure level (SPL) in six "bins" of 10 dB each. The bins are categorized as; less than 40 dB SPL, between 41 and 50 dB SPL, between 51 and 60 dB SPL, between 61 and 70 dB SPL, between 71 dB and 80 dB SPL and, greater than 80 dB SPL. The Envirogram has multiple uses. It can be used as a "patient and situation specific" counseling tool, and it can be used to help evaluate the soundscapes the patient experiences. The Envirogram can help reluctant patients evaluate their daily noise experiences, and it allows examination of unique and specific listening environments. The Envirogram facilitates a better understanding of the challenges involved with understanding speech in noise, and it can help support specific hearing aid recommendations, such as the benefits of advanced technology features across multiple soundscapes. III. Artificial Intelligence: Artificial intelligence (AI) is the simulation of human-like intelligence and/or behaviors within machines -- such as computers. AI already controls and effectively manages highly complex and intricate systems; Internet search engines, cell phone communication, automobile-based and hand-held navigation systems, chess-playing computers, computer controlled automobile driving systems, digital hearing aid amplification systems and more. All modern hearing aids incorporating AI are digital signal processing (DSP) hearing aids. However, not all modern DSP-based hearing aids possess AI. AI-based hearing aids can be thought of as "outcomes-based" digital signal processors because they select electro-acoustic parameters which assure the most favorable outcome regarding maximal speech understanding - before delivering the chosen circuit parameters to the end-user. AI-based hearing aid circuits "evaluate" the incoming signal, and -- if a better signal-to-noise ratio can be achieved via directionality, compression or noise reduction changes -- those specific changes are applied at that moment. With AI, the soundscape is evaluated and re-evaluated constantly to provide the maximum signal-to-noise ratio. The achievement of this goal is the core rationale behind the application of AI in hearing aids (Sommer, 2005). There are three basic components to Oticon's AI system, referred to as Voice Priority Processing (VPP), which work together to assure synergy via parallel processing. Flynn (2004) noted parallel processing is indeed a pre-requisite to artificial intelligence.
Artificial Intelligence in Clinical Studies: Flynn and Lunner (2005) reported on 37 adults with moderate sensorineural hearing loss, all of whom were satisfied, long-term wearers of their own previously fit premium digital signal processing hearing aids. Subjects compared their own hearing aids to Oticon's Syncro hearing aid. Results indicated consistent objective and subjective improvements while using Syncro with artificial intelligence. IV. Case Presentations: Combining the Envirogram with AI, allows the clinician to maximally address the listening needs of the individual patient, based on their hearing loss and their listening needs, as determined in their unique acoustic environment and displayed on their Envirogram. Case One:
His Envirogram is shown in Figure 2, indicating the percentage of time in each directional mode, versus input level. The overall shape of this Envirogram is common. The majority of time was spent in moderate sound environments (50 through 80 dB SPL) with less time spent in quieter and louder situations. Notice how the device recorded more time in the Split Directional mode, as compared to the Full Directional mode. At certain input levels, such as between 70 and 80 dB SPL, Syncro varies the directional mode to maximize the signal-to-noise ratio and sometimes (depending on the soundscape) the device remained in omni. These "outcomes based" settings reflect the decision making ability of the device to maximize the signal-to-noise ratio while maintaining a good sound quality and loudness balance. Without artificial intelligence, an ongoing sophisticated "balance of priorities" would not be possible and decisions based solely on a single dimension (i.e., input level) can lead to erroneous conclusions and inefficient circuit selections. Case Two:
Figure 3 demonstrates the Envirogram of an 81 year-old new hearing aid user. This woman had a mild to moderate bilateral sloping SNHL. She was fit with a set of Syncro 2 BTEs. This Envirogram was based on data collected over a one week period of time while wearing her Syncro 2 BTEs. This Envirogram reflects the percentage of time the hearing aids were worn in speech versus noise environments. This Envirogram shows a skew towards less intense acoustic environments, referred to as a "soft skew." The soft skew pattern is typical of many new hearing aid users in this age group. Her most common acoustic environments were revealed to be between 40 and 60 dB SPL. The soft skew may indicate the patient truly spends the majority of her time in soft to moderate acoustic environments. Alternatively, a soft skew may indicate the patient is not willing to try to communicate in more difficult environments. Unfortunately, the Envirogram alone cannot address this question. As is true with most audiologic data, information must be viewed within context, with due respect for test results, individual needs, history and the associated facts of the particular case. In diagnostics, the "cross check" principle (Jerger and Hayes 1976) helps assure a correct diagnosis. Likewise, the Envirogram needs to be viewed in the context of other audiologic information to maximally interpret data with respect to managing the patient, their soundscape and their listening needs. The patient was very satisfied with her Syncro 2 BTEs. When asked if she felt she was attending as many challenging acoustic environments as she would like to, she responded yes. She reported her hearing loss never kept her from going where she wanted; she just wanted to communicate more effectively in those environments. Case Three:
These Envirograms reveal significant time spent in very loud environments. The patient reported very good performance with Syncros. He compared them to his previous DSP fitting which required manually switching between directional programs. Notice how much time the device detected a "speech plus noise" signal and how much total time was spent in Split Directional or Full Directional modes. This patient had significant demands on his hearing in very loud and challenging acoustic environments. Syncro with Artificial Intelligence automatically adjusted to meet these acoustic environmental challenges. Case Four:
Figure 6 is the Envirogram of a 37 year-old male attorney. This gentleman presents with a bilateral precipitous high frequency SNHL. His hearing is essentially normal through 3000 Hz, approaches a 50 dB hearing loss at 4000, and a 70 dB loss at 6000 and 8000 Hz. He reports noise exposure including; weapons fire, chainsaws, motorcycles and he attends many live concerts. Although he reported he was able to converse with clients and judges in one-on-one situations, he was not able to hear well during trials. The patient wore a SAM device for one week before deciding which hearing aids to order. His SAM-based Envirogram revealed his typical work day was spent in quiet and medium loudness. His difficult court-room experiences were based on listening to quiet speech sounds from great distances. This patient was fit with binaural Syncro CICs and is a satisfied wearer of advanced technology hearing aids with artificial intelligence. At this time, he elects to wear his CICs exclusively during trials and important one-on-one client and business meeting Conclusions: Although professionals have previously documented hearing loss and hearing abilities, we have only recently acquired the ability to accurately record the unique acoustic environments within which each patient resides. Understanding individual patient-based soundscapes is important and useful when selecting and managing amplification for patients with hearing loss. Patient-specific soundscapes can be assessed using the free standing Sound Activity Meter (SAM) or Syncro's Activity Analyzer (AA). These two tools (SAM and AA) allow the creation of the Envirogram, which details the patient's acoustic environment with respect to; loudness level, percentages of total time and type of sound (noise, quiet, speech, or speech-in-noise). The Envirogram is a powerful new tool in the professional's armamentarium. The Envirogram reveals the patient's unique acoustic environment and encourages patient-specific counseling regarding listening needs across a multitude of dynamic listening environments. Artificial Intelligence, as applied to hearing aid amplification, allows the hearing aid wearer to experience the best possible signal-to-noise ratio at a given moment in time, across a multiple of dynamic and challenging acoustic environments. Combining the patient's personal Envirogram with Artificial Intelligence is a powerful strategy. This strategy allows the professional to document the specific needs of the patient and to create customized solutions to best address those needs. References Flynn, M.C (2004): Maximizing the Voice-to-Noise Ratio via Voice Priority Processing. The Hearing Review. April, 2004. Pages 54-59. Flynn, M.C. and Lunner, T. (2005): "Clinical Verification of a Hearing Aid with Artificial Intelligence." The Hearing Journal. February, 2005., Vol 58, No 2. Pages 34-38. Flynn, M.C. and Schmidtke, T. E. (2002). Four fitting issues for severe Flynn, M.C., Pogash R. and Schum, D (2005). Multinational clinical evaluation of Oticon Syncro. News from Oticon: Audiology Research Documentation. February 2005. Jerger, J., Hayes, D. : The Cross-Check Principle in Pediatric Audiometry. Arch Otolaryngol. Vol 102, No 10. October, 1976. Sommer, P. (2005): Artificial Intelligence -- A White Paper. Oticon A/S. |
|
||
