Temporal Association of Atrial Fibrillation With Cardiac Implanted Electronic Device Detected Heart Failure Status.

OBJECTIVES: This study sought to investigate the temporal association between changes in physiologic heart failure (HF) sensors, atrial fibrillation (AF) progression, and clinical HF in patients with cardiac resynchronization therapy implantable defibrillators (CRT-D) designed to monitor AF and HF daily. BACKGROUND: AF is a common comorbidity in HF; however, it is unclear if HF triggers AF, or vice-versa. Current implantable cardiac devices have sensors capable of quantifying HF status, which permits a greater understanding of the impact of AF on HF status and may help guide treatment. METHODS: The MultiSENSE (Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients) study collected multiple sensor data indicative of HF status in patients with CRT-D followed for up to 12 months. Patients were grouped according to their longest daily AF burden: 1) at least 24 hours of AF (HIGH AF); 2) between 6 minutes and 24 hours (MID AF); and 3) <6 minutes (NO AF). Sensor data were aligned to the first qualifying AF event or a randomly selected day for patients in the NO AF group. RESULTS: Among 869 patients with daily AF data available, 98 patients had HIGH AF, 141 patients MID AF, and 630 patients NO AF. At baseline, history of AF, N-terminal pro hormone B-type natriuretic peptide and device-measured S3 were associated with development of AF. HeartLogic index increased before AF onset (Δ HeartLogic = 9.83 ± 2.49; P < 0.001). Multivariable time-dependent Cox regression showed an increased risk for HF events following a 24-hour AF episode compared with no 24-hour AF (hazard ratio: 1.96; 95% confidence interval: 1.03-3.74). CONCLUSIONS: Device-measured HF indicators worsened before AF onset, whereas clinical HF deterioration only became apparent after AF occurred. Thus, the sensitivity of methods to ascertain AF and HF status appear to influence the direction of perceived causality. (Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients [MultiSENSE]; NCT01128166).

Ambulatory Monitoring of Heart Sounds via an Implanted Device Is Superior to Auscultation for Prediction of Heart Failure Events.

BACKGROUND: We compared the relationship between the third heart sound (S3) measured by an implantable cardiac device (devS3) and auscultation (ausS3) and evaluated their prognostic powers for predicting heart failure events (HFEs). METHODS AND RESULTS: In the MultiSENSE study, devS3 was measured daily with continuous values, whereas ausS3 was assessed at study visits with discrete grades. They were compared among patients with and without HFEs at baseline and against each other directly. Cox proportional hazard models were developed between follow-up visits and over the whole study. Simulations were performed on devS3 to match the limitations of auscultation. We studied 900 patients, of whom 106 patients experienced 192 HFEs. Two S3 sensing modalities correlated with each other, but at baseline, only devS3 differentiated patients with or without HFEs (P < 0.0001). The prognostic power of devS3 was superior to that of ausS3 both between follow-up visits (HR = 5.7, P < 0.0001, and 1.7, P = 0.047, respectively) and over the whole study (HR = 2.9, P < 0.0001, and 1.4, P = 0.216, respectively). Simulation results suggested this superiority may be attributed to continuous monitoring and to subaudible measuring capability. CONCLUSIONS: S3 measured by implantable cardiac devices has stronger prognostic power to predict episodes of future HFEs than that of auscultation.

HeartLogic Multisensor Algorithm Identifies Patients During Periods of Significantly Increased Risk of Heart Failure Events: Results From the MultiSENSE Study.

BACKGROUND: Care of heart failure (HF) patients results in a high burden on healthcare resources, and estimating prognosis is becoming increasingly important to triage resources wisely. Natriuretic peptides are recommended prognosticators in chronic HF. Our objective was to evaluate whether a multisensor HF index and alert algorithm (HeartLogic) replaces or augments current HF risk stratification. METHODS AND RESULTS: MultiSENSE (Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients) enrolled 900 patients with cardiac resynchronization therapy defibrillators enabled for collection of heart sounds, respiration, thoracic impedance, heart rate, and activity data. The HeartLogic algorithm automatically calculated a daily HF index and identified periods IN or OUT of an active alert state relative to a configurable threshold. Patients experienced 192 independently adjudicated HF events (average rate, 0.20/patient-year [pt-yr]) during 1 year of follow-up. HF event rates while IN alert was 10-fold higher than OUT of alert (0.80 versus 0.08 events/pt-yr). Combined with NT-proBNP (N-terminal pro-B-type natriuretic peptide) at enrollment (relative to 1000 pg/mL threshold, event rate was 0.42 [HIGH] versus 0.07 [LOW] events/pt-yr), substratification found the lowest risk group (LOW NT-proBNP and OUT of alert) experienced 0.02 events/pt-yr, whereas the highest risk group (HIGH NT-proBNP and IN alert) was associated with a 50-fold increased risk of an HF event (1.00 events/pt-yr) relative to the lowest risk group. CONCLUSIONS: Dynamic assessment using implantable device sensors within HeartLogic by itself or in conjunction with NT-proBNP measurements can identify time-intervals when patients are at significantly increased risk of worsening HF and potentially better triage resources to this vulnerable patient population. CLINICAL TRIAL REGISTRATION: https://www.clinicaltrials.gov. Unique identifier: NCT01128166.

Haemodynamic monitoring of cardiac status using heart sounds from an implanted cardiac device.

AIM: The aim of this study was to evaluate the haemodynamic correlates of heart sound (HS) parameters such as third HS (S3), first HS (S1), and HS-based systolic time intervals (HSTIs) from an implantable cardiac device. METHODS AND RESULTS: Two unique animal models (10 swine with myocardial ischaemia and 11 canines with pulmonary oedema) were used to evaluate haemodynamic correlates of S1, S3, and HSTIs, namely, HS-based pre-ejection period (HSPEP), HS-based ejection time (HSET), and the ratio HSPEP/HSET during acute haemodynamic perturbations. The HS was measured using implanted cardiac resynchronization therapy defibrillator devices simultaneously with haemodynamic references such as left atrial (LA) pressure and left ventricular (LV) pressure. In the ischaemia model, S1 amplitude (r = 0.76 ± 0.038; P = 0.002), HSPEP (r = -0.56 ± 0.07; P = 0.002), and HSPEP/HSET (r = -0.42 ± 0.1; P = 0.002) were significantly correlated with LV dP/dtmax . In contrast, HSET was poorly correlated with LV dP/dtmax (r = 0.14 ± 0.14; P = 0.23). In the oedema model, a physiological delayed response was observed in S3 amplitude after acute haemodynamic perturbations. After adjusting for the delay, S3 amplitude significantly correlated with LA pressure in individual animals (r = 0.71 ± 0.07; max: 0.92; min: 0.17) as well as in aggregate (r = 0.62; P < 0.001). The S3 amplitude was able to detect elevated LA pressure, defined as >25 mmHg, with a sensitivity = 58% and specificity = 90%. CONCLUSIONS: The HS parameters such as S1, S3, and HSTIs measured using implantable devices significantly correlated with haemodynamic changes in acute animal models, suggesting potential utility for remote heart failure patient monitoring.

A Multisensor Algorithm Predicts Heart Failure Events in Patients With Implanted Devices: Results From the MultiSENSE Study.

OBJECTIVES: The aim of this study was to develop and validate a device-based diagnostic algorithm to predict heart failure (HF) events. BACKGROUND: HF involves costly hospitalizations with adverse impact on patient outcomes. The authors hypothesized that an algorithm combining a diverse set of implanted device-based sensors chosen to target HF pathophysiology could detect worsening HF. METHODS: The MultiSENSE (Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients) study enrolled patients with investigational chronic ambulatory data collection via implanted cardiac resynchronization therapy defibrillators. HF events (HFEs), defined as HF admissions or unscheduled visits with intravenous treatment, were independently adjudicated. The development cohort of patients was used to construct a composite index and alert algorithm (HeartLogic) combining heart sounds, respiration, thoracic impedance, heart rate, and activity; the test cohort was sequestered for independent validation. The 2 coprimary endpoints were sensitivity to detect HFE >40% and unexplained alert rate <2 alerts per patient-year. RESULTS: Overall, 900 patients (development cohort, n = 500; test cohort, n = 400) were followed for up to 1 year. Coprimary endpoints were evaluated using 320 patient-years of follow-up data and 50 HFEs in the test cohort (72% men; mean age 66.8 ± 10.3 years; New York Heart Association functional class at enrollment: 69% in class II, 25% in class III; mean left ventricular ejection fraction 30.0 ± 11.4%). Both endpoints were significantly exceeded, with sensitivity of 70% (95% confidence interval [CI]: 55.4% to 82.1%) and an unexplained alert rate of 1.47 per patient-year (95% CI: 1.32 to 1.65). The median lead time before HFE was 34.0 days (interquartile range: 19.0 to 66.3 days). CONCLUSIONS: The HeartLogic multisensor index and alert algorithm provides a sensitive and timely predictor of impending HF decompensation. (Evaluation of Multisensor Data in Heart Failure Patients With Implanted Devices [MultiSENSE]; NCT01128166).

Feeling form: the neural basis of haptic shape perception.

The tactile perception of the shape of objects critically guides our ability to interact with them. In this review, we describe how shape information is processed as it ascends the somatosensory neuraxis of primates. At the somatosensory periphery, spatial form is represented in the spatial patterns of activation evoked across populations of mechanoreceptive afferents. In the cerebral cortex, neurons respond selectively to particular spatial features, like orientation and curvature. While feature selectivity of neurons in the earlier processing stages can be understood in terms of linear receptive field models, higher order somatosensory neurons exhibit nonlinear response properties that result in tuning for more complex geometrical features. In fact, tactile shape processing bears remarkable analogies to its visual counterpart and the two may rely on shared neural circuitry. Furthermore, one of the unique aspects of primate somatosensation is that it contains a deformable sensory sheet. Because the relative positions of cutaneous mechanoreceptors depend on the conformation of the hand, the haptic perception of three-dimensional objects requires the integration of cutaneous and proprioceptive signals, an integration that is observed throughout somatosensory cortex.

Multimodal Interactions between Proprioceptive and Cutaneous Signals in Primary Somatosensory Cortex.

The classical view of somatosensory processing holds that proprioceptive and cutaneous inputs are conveyed to cortex through segregated channels, initially synapsing in modality-specific areas 3a (proprioception) and 3b (cutaneous) of primary somatosensory cortex (SI). These areas relay their signals to areas 1 and 2 where multimodal convergence first emerges. However, proprioceptive and cutaneous maps have traditionally been characterized using unreliable stimulation tools. Here, we employed a mechanical stimulator that reliably positioned animals' hands in different postures and presented tactile stimuli with superb precision. Single-unit recordings in SI revealed that most neurons responded to cutaneous and proprioceptive stimuli, including cells in areas 3a and 3b. Multimodal responses were characterized by linear and nonlinear effects that emerged during early (∼20 ms) and latter (> 100 ms) stages of stimulus processing, respectively. These data are incompatible with the modality specificity model in SI, and provide evidence for distinct mechanisms of multimodal processing in the somatosensory system.

Feasibility of heart sounds measurements from an accelerometer within an ICD pulse generator.

OBJECTIVE: The feasibility of detecting heart sounds (HS) from an accelerometer sensor enclosed within an implantable cardioverter defibrillator (ICD) pulse generator (PG) was explored in a noninvasive pilot study on heart failure (HF) patients with audible third HS (S3). METHODS: Accelerometer circuitry enhanced for HS was incorporated into non-functional ICDs. A study was conducted on 30 HF patients and 10 normal subjects without history of cardiac disease. The devices were taped to the skin surface over both left and right pectoral regions to simulate subcutaneous implants. A lightweight reference accelerometer was taped over the cardiac apex. Waveforms were recorded simultaneously with a surface electrocardiogram for 2 minutes. Algorithms were developed to perform off-line automatic detection of HS and HS time intervals (HSTIs). RESULTS: S1, S2, and S3 vibrations were detected in all accelerometer locations for all 40 subjects, including 16 subjects without an audible S3. A substantial proportion of S3 energy was infrasonic (<20 Hz). Extending the signal bandwidth accordingly increased HS amplitudes and the ability of S3 to separate HF patients from the normal subgroup. HSTIs also separated the subgroups and were less susceptible to patient-dependent acoustic propagation properties than amplitude measures. CONCLUSION: HS, including S3 amplitude and HSTIs, may be measured using PG-embedded circuitry at implant sites without special purpose leads. Further study is warranted to determine if relative changes in heart sounds measurements can be effective in applications such as remote ambulatory monitoring of HF progression and the detection of the onset of HF decompensation.

Second-order receptive fields reveal multidigit interactions in area 3b of the macaque monkey.

Linear receptive field (RF) models of area 3b neurons reveal a three-component structure: a central excitatory region flanked by two inhibitory regions that are spatially and temporally nonoverlapping with the excitation. Previous studies also report that there is an "infield" inhibitory region throughout the neuronal RF, which is a nonlinear interactive (second order) effect whereby stimuli lagging an input to the excitatory region are suppressed. Thus linear models may be inaccurate approximations of the neurons' true RFs. In this study, we characterize the RFs of area 3b neurons, using a second-order quadratic model. Data were collected from 80 neurons of two awake, behaving macaque monkeys while a random dot pattern was scanned simultaneously across the distal pads of digits D2, 3, and 4. We used an iterative method derived from matching pursuit to identify a set of linear and nonlinear terms with significant effects on the neuronal response. For most neurons (65/80), the linear component of the quadratic RF was characterized by a single excitatory region on the dominant digit. Interactions within the dominant digit were characterized by two quadratic filters that capture the spatial aspects of the interactive infield inhibition. Interactions between the dominant (most responsive) digit and its adjacent digit(s) formed the largest class of cross-digit interactions. The results demonstrate that a significant part of area 3b responses is due to nonlinear mechanisms, and furthermore, the data support the notion that area 3b neurons have "nonclassical RF"-like input from adjacent fingers, indicating that area 3b plays a role in integrating shape inputs across digits.

Multidigit movement synergies of the human hand in an unconstrained haptic exploration task.

Although the human hand has a complex structure with many individual degrees of freedom, joint movements are correlated. Studies involving simple tasks (grasping) or skilled tasks (typing or finger spelling) have shown that a small number of combined joint motions (i.e., synergies) can account for most of the variance in observed hand postures. However, those paradigms evoked a limited set of hand postures and as such the reported correlation patterns of joint motions may be task-specific. Here, we used an unconstrained haptic exploration task to evoke a set of hand postures that is representative of most naturalistic postures during object manipulation. Principal component analysis on this set revealed that the first seven principal components capture >90% of the observed variance in hand postures. Further, we identified nine eigenvectors (or synergies) that are remarkably similar across multiple subjects and across manipulations of different sets of objects within a subject. We then determined that these synergies are used broadly by showing that they account for the changes in hand postures during other tasks. These include hand motions such as reach and grasp of objects that vary in width, curvature and angle, and skilled motions such as precision pinch. Our results demonstrate that the synergies reported here generalize across tasks, and suggest that they represent basic building blocks underlying natural human hand motions.

Automated optimal detection and classification of neural action potentials in extra-cellular recordings.

Determination of single unit spikes from multiunit spike trains plays a critical role in neurophysiological coding studies which require information about the precise timing of events underlying the neural codes that are the basis of behavior. Searching for optimal spike detection strategies has therefore been the focus of many studies over the past two decades. In this study we describe and implement an algorithm for the optimal real time detection and classification of neural spikes. The algorithm consists of three steps: noise analysis, template generation and real time detection and classification. The first step involves estimating the background noise statistics. In this step, a "cap-fitting" algorithm is used to automatically detect a spike free segment and then the mean, standard deviation and autocorrelation function of the noise are computed. The second step involves generating optimal templates of the spikes from a segment containing both noise and multiunit activity. In this step, a generalized matched filter is used to isolate a set of preliminary spikes from the noise. The first principal component of previously recorded templates is used as the deterministic signal. The preliminary spikes are then clustered in a sub-space spanned by the first three principal components to form new templates. The third step uses these templates for the real time spike detection and classification. In this step the incoming data are projected into a lower dimensional space that is designed to maximally separate the signal from the noise energy. This algorithm provides an accurate estimate of the signal to noise ratio and provides an accurate estimate of spike times and spike shapes.

Receptive field properties of the macaque second somatosensory cortex: nonlinear mechanisms underlying the representation of orientation within a finger pad.

We investigate the position invariant receptive field properties of neurons in the macaque second somatosensory (SII) cortical region. Previously we reported that many SII region neurons show orientation tuning in the center of multiple finger pads of the hand and further that the tuning is similar on different pads, which can be interpreted as position invariance. Here we study the receptive field properties of a single finger pad for a subset (n = 61) of those 928 neurons, using a motorized oriented bar that we positioned at multiple locations across the pad. We calculate both vector fields and linear receptive fields of the finger pad to characterize the receptive field properties that give rise to the tuning, and we perform an additional regression analysis to quantify linearity, invariance, or both in individual neurons. We show that orientation tuning of SII region neurons is based on a variety of mechanisms. For some neurons, the tuning is explained by simple excitatory regions, simple inhibitory regions, or some combination of these structures. However, a large fraction of the neurons (n = 20 of 61, 33%) show position invariance that is not explained well by their linear receptive fields. Finding invariance within a finger pad, coupled with the previous result of similar tuning on different pads, indicates that some SII region neurons may exhibit similar tuning throughout large regions of the hand. We hypothesize that invariant neurons play an important role in tactile form recognition.

Receptive field properties of the macaque second somatosensory cortex: representation of orientation on different finger pads.

Orientation tuning has been studied extensively in the visual system, but little is known about it in the somatosensory system. Here we investigate tuning in the second somatosensory (SII) region using a motorized stimulator that presented a small oriented bar to the 12 finger pads of digits 2-5 (D2-D5) of the macaque monkey. A subset (23%; n = 218) of the 928 SII region neurons [the same 928 neurons studied by Fitzgerald et al. (2004, 2006)] exhibited tuning, and most of these were tuned on one or two finger pads. All eight 22.5 degrees separated orientations were represented as the preferred orientation of multiple neurons, although not necessarily in equal numbers. A measure of bandwidth indicated that tuning in the SII region is sharp and is similar to the tuning observed in visual cortical areas. In addition, two-dimensional Gaussians that were fit to the tuning curves had very high r2 values, indicating that most tuning curves are both unimodal and symmetrical with respect to their preferred orientation. Most tuned neurons had additional untuned pads, although the responsiveness of these pads tended to be less than the responsiveness of tuned pads. Neurons with multiple tuned pads tended to have similar preferred orientations on their tuned pads, which can be interpreted as evidence for integration of information across fingers or as a form of positional invariance. Finally, comparison of the tuning properties showed that there are small but significant differences between the posterior, central, and anterior fields of the SII region.

Receptive field (RF) properties of the macaque second somatosensory cortex: RF size, shape, and somatotopic organization.

The detailed structure of multidigit receptive fields (RFs) in somatosensory cortical areas such as the SII region has not been investigated previously using systematically controlled stimuli. Recently (Fitzgerald et al., 2004), we showed that the SII region comprises three adjoining fields: posterior, central, and anterior. Here we characterize the RF structures of the 928 neurons that were reported in that study using a motorized oriented bar that was indented into the 12 finger pads of digits 2-5. Most (81%) of the neurons were responsive to the oriented bar stimuli, and 81% of those neurons had RFs that spanned multiple digits. Furthermore, the RFs varied greatly in size, shape, and complexity. Some RFs contained only excitatory finger pads, some contained only inhibitory pads, and some contained both types of pads. A subset of the neurons (23%) showed orientation tuning within one or more pads. The RFs spread across different digits more than within individual digits, and the responsive finger pads for a given neuron tended to cluster together within the hand. Distal and lateral finger pads were better represented than proximal and medial finger pads. Furthermore, neurons in the posterior, central, and anterior SII region fields contained different proportions of RF types. These results collectively indicate that most SII region neurons are selective for different stimulus forms either within single finger pads or across multiple pads. We hypothesize that these RFs represent the kernels underlying the representation of tactile shape.

Receptive field properties of the macaque second somatosensory cortex: evidence for multiple functional representations.

The detailed functional organization of the macaque second somatosensory cortex (SII) is not well understood. Here we report the results of a study of the functional organization of the SII hand region that combines microelectrode mapping using hand-held stimuli with single-unit recordings using a motorized, computer-controlled tactile oriented bar. The data indicate that the SII hand region extends approximately 10 mm in the anteroposterior (AP) dimension, primarily within the upper bank of the lateral sulcus. Furthermore, we find evidence that this region consists of multiple functional fields, with a central field containing neurons that are driven well by cutaneous stimuli, flanked by an anterior field and a posterior field that each contain neurons that are driven well by proprioceptive stimuli and less well by cutaneous stimuli. The anterior field extends approximately 4-5 mm AP, the central field extends approximately 3-4 mm, and the posterior field extends approximately 3 mm. Data from the motorized stimulator indicate that neurons in the central field are more responsive to oriented bars, more frequently exhibit orientation tuning, and have larger receptive fields than neurons in the anterior and posterior fields. We speculate that the three putative fields play different functional roles in tactile perception; the anterior and posterior fields process information that involves both proprioceptive and cutaneous input such as sensorimotor integration or stereognosis, whereas the central field processes primarily cutaneous information.

Spike sorting with support vector machines.

Spike sorting of neural data from single electrode recordings is a hard problem in machine learning that relies on significant input by human experts. We approach the task of learning to detect and classify spike waveforms in additive noise using two stages of large margin kernel classification and probability regression. Controlled numerical experiments using spike and noise data extracted from neural recordings indicate significant improvements in detection and classification accuracy over linear amplitude- and template-based spike sorting techniques.