False alarms in a slit-scan flow system: causes and occurrence rates. Implications and potential solutions.

A slit-scan technique was developed as a basis for an automated prescreening system for gynecologic cytology. A flow system based on this technique was fabricated and tested and results indicated that false alarms (misclassification of objects or events from normal specimens as abnormal) are the greatest remaining obstacle to development of an automated prescreening instrument. A dual view correlation system was fabricated to provide exact image-contour correlation in flow and permit precise determination of causes and occurrence rates of false alarms. This paper presents data from correlation analyses of 23 normal cytologic specimens. Major causes of false alarms and their implications to automated prescreening are discussed. A technique that would eliminate the majority of false alarms in flow is presented.     

Signal analysis of slit-scan contour data.

As a method for the preselection of alarms in gynecological cell samples, the Battelle Cytophotometry Research Group uses the slit-scan technique to obtain various cell parameters, such as the N/C ratio and the relative DNA content, from fluorescently stained cells, which are aligned one-dimensionally in the tape system designed at Battelle. The system developed at Battelle Institute analyzes all signals that exceed the background noise. As the first step in processing the slit-scan data, several threshold levels permit the separation of various artifacts. In subsequent steps, the nuclear peak is recognized, the nuclear boundaries are calculated, and seven cell parameters are determined. For the alarm detection at present only one parameter, DNA fluorescence, is used for these determinations. Visual assignment of these data to definite objects on the tape makes it possible to obtain frequency distributions of: (a) all recorded objects within the sample on the tape; (b) all signals that are classified as cells; and (c) all types of objects that preferentially cause alarms.     

Imaging in flow.

Imaging in flow has been valuable in investigating discrepancies in flow cell measurements due to cell orientation and flow dynamics. This paper discusses optical consideration in flow imaging, slit and full field imaging systems and various cell motion arresting techniques from the standpoint of image plane exposure and suitable detector choices. It concludes with an explanation of the slit-imaging techniques employed in a multidimensional slit-scan flow system and slit-scan correlation system.     

A statistical analysis of prescreening alarms in a population of normal and abnormal gynecologic specimens

A multidimensional slit-scan flow system has been developed to serve as an automated prescreening instrument for gynecological cytology. Specimens are classified abnormal based on the number of cells having elevated nuclear fluorescence (alarms). An alarm region in a bivariate histogram of nuclear fluorescence versus nuclear-to-cell-diameter ratio is defined. Alarm region probability arrays are calculated to estimate the probability that an alarm falling in a particular bin of the alarm region is either from a normal or an abnormal specimen. From these arrays, a weighted alarm index is generated. In addition, summary indices are derived that measure how the distribution of alarms in each specimen compares with the average distributions for the normal and abnormal specimen populations. These indices together with current features are evaluated with respect to their utility in specimen classification using a nonparametric classification technique known as recursive partitioning. Resulting classification trees are presented that suggest information in the distribution of alarms in the bivariate histogram. In addition, they validate the features and rules currently used for specimen classification. Recursive partitioning appears to be useful for multivariate classification and is seen as a promising technique for other applications.     

Preselection of alarms in a hybrid system for screening of cervical cancer.

In this report, a preselection of alarms in a system for automated screening of cervical cancer based on depositing the cell sample linearly as a "cell trace" on a tape and analyzing it at different decision levels with increasing complexity, and preliminary results on analyzing cervical material with this system are discussed. The "cell trace" is analyzed with the slit-scan technique. Six parameters are computed: 1) cellular diameter; 2) nuclear diameter; 3) nuclear fluorescence (acriflavin-Feulgen) as nuclear DNA; 4) cellular fluorescence; 5) nuclear to cytoplasm ratio (N/C ratio); and 6) nuclear density. At present, only nuclear fluorescence is used to define a decision boundary between normal and potentially atypical cells. Under this criteria the slit-scan analysis leaves 5% of the events in a sample that must be rechecked at a second decision level in normal cell samples. A further reduction is expected when several slit-scan parameters are used at the first decision step. All events declared suspicious will be investigated in more detail by a two dimensional image analyzing system where the fluorescence image is generated by a laser scanning system. Results obtained in preliminary experiments are discussed in this paper.     

Predicted performance of single- versus multiple-slit flow systems.

Flow systems utilizing multiple orthogonal excitation slits have been proposed as a means of reducing some types of false alarms in prescreening systems for gynecologic cytology. Such false alarms include those caused by orientation-dependent events, such as passage of binucleate or overlapping cells through the measurement region with both nuclei entering the excitation slit simultaneously. This paper presents distributions of optimal projection angles for randomly oriented nuclei passing through one, two, and three slit excitation regions. The results are used to compute observed nuclear spacing of binucleate cells and to compare performance of one, two, and three slit systems in recognition of binucleate and overlapping cells.     

Slit-scan flow cytometry: separability properties of cell features

A model is presented to compare the separability of cell populations described by features measured in low resolution slit-scanning flow systems with their separability when the features are extracted from high resolution digitized cell images. The results show that although the accuracy of the feature measurements deteriorates for increasing slit width, this is not necessarily true for the discriminatory power of the features. Depending on their original position in the high resolution feature space, the cell populations may be located even farther apart in the space of low resolution slit-scan features for reasonably small widths of the slit. The results presented with high resolution images of cells from gynecological specimens and simulated slit-scan measurements can be explained by the model. For the features nuclear DNA content and diameter the abnormal populations are shifted closer to the normal populations in the slit-scan simulations as compared to the high resolution measurements. The cell classifier errors rates are unacceptably high.     

Multidimensional slit-scan detection of bladder cancer. Preliminary clinical results

A multidimensional slit-scan flow system was developed for the automated recognition of abnormal cells derived from cancer of the uterine cervix and its precursors. It provides great sensitivity in both its ability to recognize cellular abnormality and to deal with the myriad potential causes of false alarms in an automated flow system. While its initial application was the automated recognition of the spectrum of neoplasia in gynecologic cytology samples, a preliminary study was carried out using specimens obtained from the urinary bladder. Cellular material was collected by bladder irrigation and stained with the fluorochrome acridine orange. One hundred fifty-three bladder irrigation specimens, including 115 abnormal specimens containing cells derived from neoplastic lesions of the bladder epithelium, were analyzed. For the purposes of this study, abnormal specimens from the urinary bladder included specimens containing cells derived from the following lesions of the urothelium: dysplasia (atypical hyperplasia), carcinoma-in-situ, and transitional cell carcinoma, grades 1-3. Approximately 50,000 cells were analyzed for most specimens. Of the 38 presumed normal specimens (specimens containing only normal urothelial components), four were instrument classified abnormal. For the 69 specimens containing cells derived from transitional cell carcinoma, grade 1, 1-2, 2, 66 were correctly classified as abnormal while three were classified as normal.(ABSTRACT TRUNCATED AT 250 WORDS)     

Ambient noise increases missed detections in nestling birds

Ambient noise can mask acoustic cues, making their detection and discrimination difficult for receivers. This can result in two types of error: missed detections, when receivers fail to respond to the appropriate cues, and false alarms, when they respond to inappropriate cues. Nestling birds are error-prone, sometimes failing to beg when parents arrive with food (committing missed detections) or begging in response to stimuli other than a parent's arrival (committing false alarms). Here, we ask whether the frequency of these errors by nestling tree swallows (Tachycineta bicolor) increases in the presence of noise. We found that nestlings exposed to noise had more missed detections than their unexposed counterparts. We also found that false alarms remained low overall and did not differ significantly between noise and quiet treatments. Our results suggest that nestlings living in noisy environments may be less responsive to their parents than nestlings in quieter environments.     

A multidimensional slit-scan flow system.

A new slit-scan type flow system is described which provides three (X, Y, and Z) orthogonal one-dimensional projections of cell fluorescence. A photomultiplier tube and two semiconductor array detectors are used to obtain the three slit-scan contours from cells traversing a single fluorescence excitation beam. A high speed, dedicated preprocessor analyzes the three contours in parallel, extracting certain features useful for rejecting cells from which an accurate measurement of nuclear fluorescence cannot be obtain. Contour data is buffered and transferred to a PDP-11/40 computer where nuclear fluorescence is measured and cells are classified. It is anticipated that this new instrument will provide a significant reduction in false alarm rate when applied to prescreening of gynecologic cytology specimens.     

Fork-tailed drongos use deceptive mimicked alarm calls to steal food

Despite the prevalence of vocal mimicry in animals, few functions for this behaviour have been shown. I propose a novel hypothesis that false mimicked alarm calls could be used deceptively to scare other species and steal their food. Studies have previously suggested that animals use their own species-specific alarm calls to steal food. However none have shown conclusively that these false alarms are deceptive, or that mimicked alarm calls are used in this manner. Here, I show that wild fork-tailed drongos (Dicrurus adsimilis) make both drongo-specific and mimicked false alarm calls when watching target species handling food, in response to which targets flee to cover abandoning their food. The drongo-specific and mimicked calls made in false alarms were structurally indistinguishable from calls made during true alarms at predators by drongos and other species. Furthermore, I demonstrate by playback experiments that two of these species, meerkats (Suricata suricatta) and pied babblers (Turdoides bicolor), are deceived by both drongo-specific and mimicked false alarm calls. These results provide the first conclusive evidence that false alarm calls are deceptive and demonstrate a novel function for vocal mimicry. This work also provides valuable insight into the benefits of deploying variable mimetic signals in deceptive communication.     

Negative feedback from maternal signals reduces false alarms by collectively signalling offspring

Within animal groups, individuals can learn of a predator's approach by attending to the behaviour of others. This use of social information increases an individual's perceptual range, but can also lead to the propagation of false alarms. Error copying is especially likely in species that signal collectively, because the coordination required for collective displays relies heavily on social information. Recent evidence suggests that collective behaviour in animals is, in part, regulated by negative feedback. Negative feedback may reduce false alarms by collectively signalling animals, but this possibility has not yet been tested. We tested the hypothesis that negative feedback increases the accuracy of collective signalling by reducing the production of false alarms. In the treehopper Umbonia crassicornis, clustered offspring produce collective signals during predator attacks, advertising the predator's location to the defending mother. Mothers signal after evicting the predator, and we show that this maternal communication reduces false alarms by offspring. We suggest that maternal signals elevate offspring signalling thresholds. This is, to our knowledge, the first study to show that negative feedback can reduce false alarms by collectively behaving groups.     

False alarms and information transmission in grouping animals

A key benefit of grouping in prey species is access to social information, including information about the presence of predators. Larger groups of prey animals respond both sooner and at greater distances from predators, increasing the likelihood that group members will successfully avoid capture. However, identifying predators in complex environments is a difficult task, and false alarms (alarm behaviours without genuine threat) appear surprisingly frequent across a range of taxa including insects, amphibians, fish, mammals, and birds. In some bird flocks, false alarms have been recorded to substantially outnumber true alarms. False alarms can be costly in terms of both the energetic costs of producing alarm behaviours as well as lost opportunity costs (e.g. abandoning a feeding patch which was in fact safe, losing sleep if an animal is resting/roosting, or losing mating opportunities). Models have shown that false alarms may be a substantial but underappreciated cost of group living, introducing an inherent risk to using social information and a vulnerability to the propagation of false information. This review will focus on false alarms, introducing a two-stage framework to categorise the different factors hypothesised to influence the propensity of animal groups to produce false alarms. A number of factors may affect false alarm rate, and this new framework splits these factors into two core processing stages: (i) individual perception and response; and (ii) group processing of predator information. In the first stage, individuals in the group monitor the environment for predator cues and respond. The factors highlighted in this stage influence the likelihood that an individual will misclassify stimuli and produce a false alarm (e.g. lower light levels can make predator identification more difficult and false alarms more common). In the second stage, alarm information from individuals is processed by the group. The factors highlighted in this stage influence the likelihood of alarm information being copied by group members and propagated through the group (e.g. some animals implement group processing mechanisms that regulate the spread of behavioural responses such as consensus decision making through the quorum response). This review follows the structure of this new framework, focussing on the causes of false alarms, factors that influence false alarm rate, the transmission of alarm information through animal groups, mechanisms to mitigate the spread of false alarms, and the consequences of false alarms.     

Filtering intrusion detection alarms

A Network Intrusion Detection System (NIDS) is an alarm system for networks. NIDS monitors all network actions and generates alarms when it detects suspicious or malicious attempts. A false positive alarm is generated when the NIDS misclassifies a normal action in the network as an attack. We present a data mining technique to assist network administrators to analyze and reduce false positive alarms that are produced by a NIDS. Our data mining technique is based on a Growing Hierarchical Self-Organizing Map (GHSOM) that adjusts its architecture during an unsupervised training process according to the characteristics of the input alarm data. GHSOM clusters these alarms in a way that supports network administrators in making decisions about true and false alarms. Our empirical results show that our technique is effective for real-world intrusion data.     

Invariance and selectivity in the ventral visual pathway.

Pattern recognition systems that are invariant to shape, pose, lighting and texture are never sufficiently selective; they suffer a high rate of "false alarms". How are biological vision systems both invariant and selective? Specifically, how are proper arrangements of sub-patterns distinguished from the chance arrangements that defeat selectivity in artificial systems? The answer may lie in the nonlinear dynamics that characterize complex and other invariant cell types: these cells are temporarily more receptive to some inputs than to others (functional connectivity). One consequence is that pairs of such cells with overlapping receptive fields will possess a related property that might be termed functional common input. Functional common input would induce high correlation exactly when there is a match in the sub-patterns appearing in the overlapping receptive fields. These correlations, possibly expressed as a partial and highly local synchrony, would preserve the selectivity otherwise lost to invariance.     

Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients

Purpose

Physiologic monitors are plagued with alarms that create a cacophony of sounds and visual alerts causing “alarm fatigue” which creates an unsafe patient environment because a life-threatening event may be missed in this milieu of sensory overload. Using a state-of-the-art technology acquisition infrastructure, all monitor data including 7 ECG leads, all pressure, SpO2, and respiration waveforms as well as user settings and alarms were stored on 461 adults treated in intensive care units. Using a well-defined alarm annotation protocol, nurse scientists with 95% inter-rater reliability annotated 12,671 arrhythmia alarms.

Results

A total of 2,558,760 unique alarms occurred in the 31-day study period: arrhythmia, 1,154,201; parameter, 612,927; technical, 791,632. There were 381,560 audible alarms for an audible alarm burden of 187/bed/day. 88.8% of the 12,671 annotated arrhythmia alarms were false positives. Conditions causing excessive alarms included inappropriate alarm settings, persistent atrial fibrillation, and non-actionable events such as PVC''s and brief spikes in ST segments. Low amplitude QRS complexes in some, but not all available ECG leads caused undercounting and false arrhythmia alarms. Wide QRS complexes due to bundle branch block or ventricular pacemaker rhythm caused false alarms. 93% of the 168 true ventricular tachycardia alarms were not sustained long enough to warrant treatment.

Discussion

The excessive number of physiologic monitor alarms is a complex interplay of inappropriate user settings, patient conditions, and algorithm deficiencies. Device solutions should focus on use of all available ECG leads to identify non-artifact leads and leads with adequate QRS amplitude. Devices should provide prompts to aide in more appropriate tailoring of alarm settings to individual patients. Atrial fibrillation alarms should be limited to new onset and termination of the arrhythmia and delays for ST-segment and other parameter alarms should be configurable. Because computer devices are more reliable than humans, an opportunity exists to improve physiologic monitoring and reduce alarm fatigue.     

Slit-scan cytofluorometry. Basis for automated prescreening of urinary tract cytology.

A study was undertaken to assess the applicability of the slit-scan technique to automated prescreening of urinary tract cytology. Cells from voided and catheterized urines were stained with acridine orange and measured on a static cell slit-scan cytofluorometer. Analysis of data from the specimens indicates that nuclear fluorescence alone appears adequate for recognition of abnormal specimens. Remaining problems in the automation of urinary tract cytology prescreening are discussed.     

Image cytometry in automated cervical screening

Severe restrictions with regard to false negative rates have played a major role in the development of the LEYden Television Analysis System (LEYTAS). The present paper describes a test with a continuous series of 1500 cervical samples illustrating the accuracy of LEYTAS in a fully automated screening procedure using cell selection transformations and artefact rejection procedures. Specimen classification with a cut-off at greater than 0.3% alarms (= percentage of automatically selected objects per epithelial cells) and greater than 10 alarms, results in a false negative rate (FNR) of 0.3% (1 case out of 321 cases with severe dysplasia or more serious lesions), a false positive rate (FPR) of 13% (663 negative cases) and a rejection rate of 2.7%. Besides a machine classification, LEYTAS offers a second, machine-interaction classification of those preparations which have been declared positive by the machine. Machine-interaction involves visual evaluation of the stored images of the detected objects (alarms) and reduces the FPR from 13 to 8%. Statistical tests further demonstrate the significance of the screening results. Presently the main drawback for routine use of automated screening with LEYTAS seems to be the time consuming preparation procedure, since instrumentation has now been updated to a new, fast and user-friendly version of LEYTAS.     

System for flow sorting chromosomes on the basis of pulse shape

Sorting on the basis of the complex features resolved by chromosome slit-scan analysis requires rapid and flexible pulse shape acquisition and processing for determining sort decisions before droplet breakoff. Fluorescence scans of chromosome morphology contain centromeric index and banding information suitable for chromosome classification, but these scans are often characterized by variability in length and height and require sophisticated data processing procedures for identification. Setting sort criteria on such complex morphological data requires digitization and subsequent computation by an algorithm tolerant of variations in overall pulse shape. We demonstrate here the capability to sort individual chromosomes based on their morphological features measured by slit-scan flow cytometry. To do this we have constructed a sort controller capable of acquiring an 128 byte chromosome waveform and executing a series of numerical computations resulting in an area-based centromeric index sort decision in less than 2 ms. The system is configured in a NOVIX microprocessor, programmed in FORTH, and interfaced to a slit-scan flow cytometer data acquisition system. An advantage of this configuration is direct control over the machine state during program execution for minimal processing time. Examples of flow sorted chromosomes are shown with their corresponding fluorescence pulse shapes.     

Application of motion analysis system in pre-impact fall detection

The purpose of this study is to investigate unique features of body segments in fall and activities of daily living (ADL) to make automatic detection of fall in its descending phase before the impact. Thus, fall-related injuries can be prevented or reduced by deploying feedback systems before the impact. In this study, the authors propose the following hypothesis: (1) thigh segment normally does not go beyond certain threshold angle to forward and sideways directions in ADL and (2) even if it does, the angular characteristics measured at torso and thigh differ from one another in ADL whereas in the case of fall, they become congruent. These two factors can be used to distinguish fall from ADL in its inception. Vicon 3-D motion analysis system was used in this study. High level of correlation between thigh and torso segments (corr > 0.99) was found for fall activities and low correlation coefficients (mean corr for lateral movements is 0.2338 and for sagittal movements is -0.665) were observed in ADL. By applying the hypothesis, all simulated falls could be detected with no false alarms and around 700ms lead-time before the impact was achieved in pre-impact fall detection. It is the longest lead-time obtained so far in pre-impact fall detection.