Jean-François Motsch

Objective detection of brainstem auditory evoked potentials with a priori information from higher presentation levels

This paper describes a brainstem auditory evoked potentials (BAEPs) detection method based on supervised pattern recognition. A previously used pattern recognition technique relying on cross-correlation with a template was modified in order to include a priori information allowing detection accuracy. Reference is made to the patients audiogram and to the latency-intensity (LI) curve with respect to physiological mechanisms. Flexible and adaptive constraints are introduced in the optimization procedure by means of eight rules. Several data samples were used in this study. The determination of parameters was performed through 270 BAEPs from 20 subjects with normal and high audiometric thresholds and through additional BAEPs from 123 normal ears and 14 ears showing prominent wave VI BAEPs. The evaluation of the detection performance was performed in two steps: first, the sensitivity, specificity and accuracy were estimated using 283 BAEPs from 20 subjects showing normal and high audiometric thresholds and secondly, the sensitivity, specificity and accuracy of the detection and the accuracy of the response threshold were estimated using 213 BAEPs from 18 patients in clinic. Taking into account some a priori information, the accuracy in BAEPs detection was enhanced from 76 to 90%. The patient response thresholds were determined with a mean error of 5 dB and a standard deviation error of 8.3 dB. Results were obtained using experimental data; therefore, they are promising for routine use in clinic.

Source level estimation of two blue whale subspecies in southwestern Indian Ocean

Blue whales produce intense, stereotypic low frequency calls that are particularly well suited for transmission over long distances. Because these calls vary geographically, they can be used to gain insight into subspecies distribution. In the Southwestern Indian Ocean, acoustic data from a triad of calibrated hydrophones maintained by the International Monitoring System provided data on blue whale calls from two subspecies: Antarctic and pygmy blue whales. Using time difference of arrival and least-squares hyperbolic methods, the range and location of calling whales were determined. By using received level of calls and propagation modeling, call source levels of both subspecies were estimated. The average call source level was estimated to 179+/-5 dB re 1 microPa(rms) at 1 m over the 17-30 Hz band for Antarctic blue whale and 174+/-1 dB re 1 microPa(rms) at 1 m over the 17-50 Hz band for pygmy blue whale. According to previous estimates, slight variations in the source level could be due to inter-individual differences, inter-subspecies variations and the calculation method. These are the first reported source level estimations for blue whales in the Indian Ocean. Such data are critical to estimate detection ranges of calling blue whales.

Male sperm whale acoustic behavior observed from multipaths at a single hydrophone

Sperm whales generate transient sounds (clicks) when foraging. These clicks have been described as echolocation sounds, a result of having measured the source level and the directionality of these signals and having extrapolated results from biosonar tests made on some small odontocetes. The authors propose a passive acoustic technique requiring only one hydrophone to investigate the acoustic behavior of free-ranging sperm whales. They estimate whale pitch angles from the multipath distribution of click energy. They emphasize the close bond between the sperm whales physical and acoustic activity, leading to the hypothesis that sperm whales might, like some small odontocetes, control click level and rhythm. An echolocation model estimating the range of the sperm whales targets from the interclick interval is computed and tested during different stages of the whales dive. Such a hypothesis on the echolocation process would indicate that sperm whales echolocate their prey layer when initiating their dives and follow a methodic technique when foraging.

Measuring the off-axis angle and the rotational movements of phonating sperm whales using a single hydrophone

The common use of the bent-horn model of the sperm whale sound generator describes sperm whale clicks as the pulse series {p0, p1, p2, p3,...}. Clicks, however, deviate from this standard when recorded using off-axis hydrophones. The existence of additional pulses within the {p0, p1, p2, p3, ...} series can be explained still using the bent-horn model. Multiple reflections on the whales frontal and distal sacs of the p0 pulse lead to additional sets of pulses detectable using a farfield, off-axis hydrophone. The travel times of some of these additional pulses depend on the whales orientation. The authors propose a method to estimate the off-axis angle of sperm whale clicks. They also propose a method to determine the nature of the movement (if it is pitch, yaw, or roll) of phonating sperm whales. The application of both methods requires the measurement of the travel time differences between pulses composing a sperm whale click. They lead, using a simple apparatus consisting of a single hydrophone at an unknown depth, to new measurements of the underwater movements of sperm whales. Using these methods shows that sperm whales would methodically scan seawater while searching for prey, by making periodic pitch and yaw movements in sync with their acoustic activity.

Non-Stationary Time-Series Segmentation Based on the Schur Prediction Error Analysis

This paper proposes a non-stationary time-series segmentation method based on the analysis of the forward prediction error issued from the adaptive Schur orthogonal signal parameterisation. There is no a priori information about the analysed signal thus this method can be easily adapted to a large family of different types of signals for which two different stochastic processes are present. In this paper we set out some of the advantages of the adaptive Schur filter in deducing the presence of different non-stationary transient or long-term events leading to the signal segmentation. For each sample, the adaptive Schur algorithm calculates the optimal second-order solution for the signal prediction resulting in a set of time-varying model parameters (inter alia forward prediction error). We define the likelihood ratio (LR) test based on the Schur forward prediction error that is evaluated at each sample, thus giving excellent time-reaction properties. The LR test allows us to effectively partition the analysed time-series into homogeneous segments by considering its second-order statistics which are tracked adaptively by the Schur filter. The results performed by applying the proposed method to simulated signals are shown to verify its high performance

Computer-assisted ABR Interpretation using the Automatic Construction of the Latency-Intensity Curve: Interpretatión asistida por computadora del ABR utilizando la Constructión Automática de la Curva Latencia-Intensidad

In this paper, we present a new computerised technique for the automatic construction of the latency-intensity curve (LI curve). We take a pattern recognition approach determined by a priori information. We use knowledge gained from the audiogram and from physiological considerations. Therefore, we consider all recordings at different intensities as well as results from the extraction of a single auditory brainstem response (ABR) at a given stimulus intensity. We tested our method successfully: it allows us to prevent misrecognition errors in response detection or in latency measurements. Automatic recognition of the waves and recognition by the ear, nose and throat (ENT) specialist coincided in at least 90 per cent of cases. For wave V, the average deviation between the response thresholds given by our automatic recognition algorithm and those given by the ENT specialist was 5 dB, and the average deviation of the latencies was 0.05 ms. En este trabajo, presentamos una nueva técnica computarizada para la constructión automática de la curva de latencia-intensidad (Curva LI). Tomamos un enfoque de reconocimiento de patrones, supervisado por informatión a priori, y hacemos referenda al conocimiento obtenido del audiograma y también a otras consideraciones fisiológicas. Por lo tanto, consideramos todos los registros a difcrentes intensidades, al igual que los resultados dc la extractión de una respuesta auditiva del tallo cerebral (ABR) única, a una intensidad dada del estímulo. Evaluamos con éxito nuestro método, lo que nos permite evitar errores de reconocimiento en la identificatión de respuestas o en las medidas de latencia. El reconocimiento automático de las ondas y el reconocimiento por parte del especialista en oídos, nariz y garganta (ORL) coincidieron en el 90 por ciento de los casos. Para la onda V, la desviación promedio entre el umbral de respuesta dado por nuestro algoritmo de reconocimiento automático y la brindada por el especialista en ORL fue dc 5 dB, y la desviacion promedio de las latencias fue de 0.05 ms.

Depth/range localization of diving sperm whales using passive acoustics on a single hydrophone disregarding seafloor reflections

Depth and range of sound sources can be estimated using a single hydrophone. Such a passive acoustic technique requires the detection of direct path transmitted, sea surface, and seafloor reflected source signals, so as to measure their time of arrival differences (TOADs). Sperm whales almost continuously emit powerful, directional echolocation sounds (usual clicks) when diving. Sperm whales often dive in deep water, and click seafloor reflections are usually well detected only at the beginning of the dive. Surface echoes may be detected during the entire dive. If the measurement of the surface reflection TOAD of a single click is not enough for estimating the depth/range of the sperm whale at the time when this click was emitted, the joint consideration of delays emitted during the whole dive may provide this estimation. Such delays are indeed the measurements of a single phenomenon: the underwater movements of a single‐clicking sperm whale. One can merge such data using a Bayesian technique, as well as ...

Adaptative Schur algorithm dedicated to underwater transient signal processing

The algorithm proposed by Lee and Morf [IEEE Transactions on Circuits and Systems 28(6) (1981)] which stems from the method defined by Schur [Operator Theory: Advances & Application, Vol. 18 (Birk‐Verlag, 1986)] has acquired a new significance [Zarzycki, Journal of Multidimensional Systems & Signal Processing (Kluwer Academic, 2004)], due to its performances and particularly due to its applications in real time, made possible by the speed of processors available nowadays. Based on the innovations filter principle, Schur’s proposal models the signal by calculating reflection coefficients, describing entirely the second‐order signal. The reflection coefficients can be simply transformed to the AR coefficients, from which one derives the time‐frequency representation. We compare performances of this approach with other time‐frequency representations commonly used in the signal processing (spectrogram, AR, wavelet transform); we subsequently present the results obtained for transitory underwater acoustic signals, which our laboratory is investigating. The Lee and Morf algorithm offers an excellent tracking of the signal’s characteristics and allows us to systematically detect transitory signals. This is particularly pertinent to segmentation problems relating to the application of underwater acoustics. The robustness of the Schur detector and a resolution of the Schur time‐frequency representation support the resurgence of the Schur algorithm.

Acoustic localization of two distinct blue whale (Balaenoptera musculus) subspecies in the South‐West Indian Ocean

Analysis of one year of acoustic signal recordings from the five permanent autonomous hydrophones of the International Monitoring System in the South‐West Indian Ocean reveals low frequency with high intensity calls produced by two blue whale subspecies. The “Antarctic” or “true” blue whale (B. m. intermedia) calls and the “Madagascar‐type” Pygmy blue whale calls (B. m. brevicauda) were automatically detected through the matched filtering method. The potential movements were investigated by using the time difference of arrival (TDOA) of calls to assess the bearing of the sound source. The fully range dependent parabolic equation code (RAM ‐ range‐dependent acoustic model) and the PMCC code (progressive multi‐channel correlation) are applied to estimate the range between our system and the vocalising animals. Our results show that (1) the variation of call number revealed two distinct patterns of seasonal whale occurrences and (2) the distances from the hydrophones to the blue whales reached up to 50 km. Tracking whales is possible when whales are concentrated of the hydrophone array.

Broken line 3‐dimensional sperm whale diving trajectory reconstruction using passive acoustics on a single hydrophone

Sperm whales make deep dives to hunt. A dive, lasting 45 minutes on average, is composed of a vertical descent to the prey layer depth, the properly so called hunt (at a quasi constant depth) inside the prey layer, and a vertical ascent back to the sea surface. Sperm whales make series of echolocation signals (clicks) during the two first stages. The sea surface/bottom click echo detection and delay measurements then make possible the sperm whale range/depth estimation during these stages, by passively using a single hydrophone. The vertical, rectilinear sperm whale trajectory during the first stage is unambiguously estimated from the echo delays. The sperm whale trajectory can also be reconstructed during the second stage, from the sperm whale range variations only, even when not detecting sea bottom click echoes. These range variations strongly suggest the sperm whale trajectory to be a broken line (e.g., a 2‐piece line, 600 m straight ahead, 85 degree bend, 1000 m straight ahead). Assuming a vertical s...