Search Results

The Falcon HST is an EEG and respiratory signal recorder. The device is intended for use by adult patients in the home or clinical environment, under the direction of a qualified healthcare practitioner, to aid in the diagnosis of sleep disorders.

The Falcon HST comprises hardware and software which provide separate parameters for recording, review, and analysis of collected and stored physiological parameters, including EEG, EOG, ECG and respiratory signals, which are then used as an aid in the diagnosis of respiratory and/or cardiac related sleep disorders by qualified physicians.

The Falcon system consists of the main unit and the charging cradle. The main unit is a small device that is worn on the patient's chest over clothing. It is equipped with a touch-screen LCD and contains various channel inputs such as for the inductive plethysmography bands, and electrodes. The Falcon charging cradle is used to charge the device, as well as provide a USB interface for transferring study data to the PC.

The manufacturer, Compumedics Limited, demonstrates the substantial equivalence of the Falcon HST to its predicate devices for aid in the diagnosis of sleep disorders. The acceptance criteria and the study that proves the device meets the acceptance criteria are described below.

Acceptance Criteria and Reported Device Performance

The acceptance criteria for the Falcon HST are based on establishing substantial equivalence to its predicate devices, the Zmachine Synergy (K172986) and Zmachine DT-100 (K101830). The performance of the Falcon HST was deemed acceptable if its capabilities for recording EEG, respiratory effort, respiratory airflow, body position, and pulse oximetry, as well as its software's ability to produce AHI and sleep staging results, were found to be substantially equivalent to the predicate devices and gold standard polysomnography (PSG) data.

Study Details:

1. Sample sizes used for the test set and the data provenance:

   • Bench Testing (Side-by-Side Comparison): The text does not specify the exact number of devices or data points used for the side-by-side bench comparison tests between the Falcon HST and Zmachine Synergy for EEG input, respiratory effort, respiratory airflow, body position, and pulse oximetry. The description implies at least one of each device was used, subject to repeated measurements or simulated inputs.
   • Clinical Performance Testing (Profusion PSG Software): The text does not provide a specific sample size for the clinical performance testing used to validate the software's ability to calculate AHI and sleep staging.
   • Data Provenance: The document does not explicitly state the country of origin or whether the data for the mentioned tests was retrospective or prospective. For bench testing, it involved simulated inputs or direct comparison against predicate devices. For clinical performance testing of the software, it's compared against "gold standard polysomnography data," implying real patient data.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience):

   • For the clinical performance testing of the Profusion PSG software, the ground truth for AHI and sleep staging was established through "expert review of gold standard polysomnography data." The number of experts and their specific qualifications (e.g., years of experience, specific certifications) are not specified in the provided text.

3. Adjudication method (e.g. 2+1, 3+1, none) for the test set:

   • The document does not specify an adjudication method for establishing ground truth, particularly for the clinical performance testing where expert review was used.

4. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:

   • The provided text does not mention an MRMC comparative effectiveness study involving human readers or any AI assistance. The study focuses on the device's equivalence to existing technology and the accuracy of its software against expert-reviewed data, not on human reader performance with or without AI assistance. The Falcon HST is an EEG and respiratory signal recorder, and its software is used to aid in the diagnosis by processing these signals, not primarily as an AI assistance tool for human interpretation in the context of what would typically be considered an MRMC study for AI.

5. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:

   • For the hardware components (EEG, respiratory effort, airflow, body position, pulse oximetry), standalone performance testing was conducted by comparing the Falcon HST's output directly against that of the predicate devices or against angular references/simulated inputs.
   • For the Profusion PSG software 5.1, its ability to calculate AHI and sleep staging was validated by comparing its outputs directly against "expert review of gold standard polysomnography data." This indicates a standalone performance evaluation of the algorithm's output against established ground truth, effectively without human-in-the-loop for the algorithm's calculation step itself. The device is intended "to aid in the diagnosis," implying that its output will be reviewed by a human practitioner.

6. The type of ground truth used (expert consensus, pathology, outcomes data, etc):

   • For testing the accuracy of AHI and sleep staging calculations by the Profusion PSG software, the ground truth used was expert review of gold standard polysomnography data.
   • For the bench testing of individual physiological parameters (EEG, respiratory effort, airflow, body position, pulse oximetry), the ground truth was established by comparison to the predicate device (Zmachine Synergy) or by using controlled simulated inputs and angular references.

7. The sample size for the training set:

   • The document does not provide any information regarding a training set size. This might be because the device's algorithms or software features (like sleep staging) may have been developed and validated previously, or the submission focuses on demonstrating equivalence to established technologies rather than novel algorithm training. The software, Profusion PSG software 5.1, is mentioned to be identical to versions previously cleared (K072201 and K093223), suggesting its core functionality and training (if any) happened prior to this submission.

8. How the ground truth for the training set was established:

   • As no information about a training set is provided, how its ground truth was established is not detailed in the document.

Ask a specific question about this device

The Nurochek System is intended for prescription use in healthcare facilities or clinical research environments for subjects ages 14 years and older. The Nurochek System is indicated for the generation of visual evoked potentials (VEPs) and to acquire, transmit, display and store electroencephalograms (EEGs) during the generation of VEPs. The Nurochek System only acquires and displays physiological signals: no claims are being made for use as a diagnostic criterion or for the analysis of the acquired signals with respect to the accuracy, precision and reliability.

The Nurochek System combines hardware, firmware and software to generate and acquire physiological signals, specifically, VEPs. These VEPs are generated by a visual stimulus delivered through the Nurochek headset worn by the subject. This visual stimulus is a short-duration flash of white light. The Nurochek headset acquires the VEPs from the rear of the head and transmits the resulting EEG to the Nurochek software application to be displayed to the user and stored. These acquired signals are intended to be analyzed by a Physician. The Nurochek System operates on the principles of generating VEPs via photic simulation and acquiring the VEPs via EEG. Photic stimulation is provided through short-duration flashes of white light from multiple LEDs located in the front of the headset to direct the stimulus into the subject's eyes. The VEPs are acquired by an EEG comprising of a total of 5 electrode interfaces with hydrophilic foam cylinders saturated with saline solution to provide electrical contact to the subject's scalp. A Bluetooth receiver and transmitter located within the Nurochek headset allows it to communicate with and be controlled by the Nurochek software application. The Nurochek software application provides a graphical user interface which allows: Collection of the subject details and consent, Initiation of a study and tracking of patient information, Acquisition and transmission of signals wirelessly to and from the headset, Display of the contact quality of electrodes to the subject's scalp, Recording, processing and display of EEG signals received from the headset, and Manage previous EEG recordings of VEPs.

The provided document, a 510(k) Premarket Notification for the Nurochek System, focuses on demonstrating substantial equivalence to predicate devices rather than proving the device meets specific acceptance criteria through a detailed clinical study with performance metrics. While it mentions non-clinical performance data and a small clinical study, it does not explicitly provide a table of acceptance criteria with corresponding performance data in the format requested.

The document states that "The Nurochek System only acquires and displays physiological signals: no claims are being made for use as a diagnostic criterion or for the analysis of the acquired signals with respect to the accuracy, precision and reliability." This statement is critical as it limits the scope of claims and thus the type of performance data required for acceptance. The study described focuses on functional equivalence in signal acquisition rather than diagnostic accuracy.

Despite this, I can extract information related to the device's performance based on the functional capabilities and a small clinical study mentioned.

Here's an attempt to answer your request based on the provided text, acknowledging the limitations in scope for a device that "only acquires and displays physiological signals and makes no claims in relation to diagnoses."

1. Table of acceptance criteria and the reported device performance:

Since the document defines the device as one that "only acquires and displays physiological signals" and makes "no claims ... for the analysis of the acquired signals with respect to the accuracy, precision and reliability," the acceptance criteria are not in terms of diagnostic performance (e.g., sensitivity, specificity). Instead, they are related to the functional equivalence of signal acquisition and display compared to a benchmark device, and safety/technical compliance.

• Dynamic Range: Nurochek: +/- 187,500 μV (superior) vs. Predicate: +/- 1000μV.
• Resolution: Nurochek: 0.02μV (superior/more accurate) vs. Predicate: 0.03μV.
• Peak to Peak Noise: Nurochek: 1.97μV (typical) (lower/superior) vs. Predicate: 3.7μV (typical).
• Common Mode Rejection Ratio: Nurochek: 110 dB vs. Predicate: 110 dB (same).
• Input Impedance: Nurochek: 1GOhm vs. Predicate: 100GOhm (Both exceed OSET recommendation of 10 MOhm).
• Impedance Check Functionality: Both have impedance check. |

2. Sample size used for the test set and the data provenance:

• Test Set Sample Size: 20 participants for the clinical study comparing SSVEP detection.
• Data Provenance: The document does not explicitly state the country of origin for the clinical study data or if it was retrospective or prospective. Given the submitter (Cryptych Pty Ltd) is based in North Sydney, NSW, Australia, and the submission is to the FDA, it's possible the study was conducted in Australia, but this is not confirmed. It was a "clinical study," which typically implies prospective data collection, but this is also not definitively stated as "prospective."

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• The document mentions using the "clinical EEG Compumedics Grael EEG as a reference device and benchmark" for the clinical study. It does not mention experts establishing a "ground truth" in the traditional sense of labeling data (e.g., for diagnostic AI). The comparison was presumably based on the raw or processed signals from the two devices themselves.
• Given the device "only acquires and displays physiological signals" and makes "no claims ... for the analysis of the acquired signals with respect to the accuracy, precision and reliability," the concept of "ground truth" for diagnostic accuracy is not applicable as per the device's intended use. The "ground truth" essentially refers to the output of the benchmark medical device.

4. Adjudication method for the test set:

• Not applicable as the study involved comparing raw physiological signal acquisition between two devices, not interpretation or diagnosis requiring expert consensus/adjudication. The statement "both systems functioned identically in their ability to detect SSVEPs" suggests a direct comparison of the acquired signals or derived metrics.

5. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:

• No, a multi-reader multi-case (MRMC) comparative effectiveness study was not performed.
• This device is not an AI-assisted diagnostic tool, but rather a device for acquiring and displaying physiological signals (VEPs and EEGs). Therefore, a study on human reader improvement with AI assistance is not relevant to its stated indications for use.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:

• The non-clinical performance data (electrical safety, light safety, biocompatibility, cleaning, mechanical, firmware/software) represent standalone testing of the device's technical aspects and capabilities.
• The clinical study was a comparison of signal acquisition between the Nurochek System and a reference EEG device; it wasn't an "algorithm-only" performance study in an AI context, but rather a verification of the physiological signal acquisition system.

7. The type of ground truth used:

• The "ground truth" for the clinical study was the performance of a cleared predicate/reference medical device (Compumedics Grael EEG) in detecting SSVEPs. This falls under the category of "comparison to a reference standard medical device."
• For the non-clinical tests, the "ground truth" was compliance with established international and national standards (e.g., IEC, ISO, ANSI, FDA guidances).

8. The sample size for the training set:

• The document implies that the device "only acquires and displays physiological signals," and the software manages and displays these. It does not describe a machine learning model that would require a "training set" in the sense of supervised learning for classification or prediction. Therefore, the concept of a training set for an AI model is not applicable here. The software validation refers to standard software engineering verification and validation activities (IEC 62304), not AI model training.

9. How the ground truth for the training set was established:

• As a training set for an AI model is not applicable, this question is not relevant.

Ask a specific question about this device