A careful biomedicine? Generalization and abstraction in RCTs

This article takes up biomedical and public health concerns about the difficulty of generalizing or extrapolating measurements of efficacy produced by the method of the randomized control trial (RCT) to wider populations. While explanations for the difficulty may be deduced from social studies of science that reveal the contingent and situated nature of trial findings, new conceptual tools are required to allow for the practical value associated with the possibility of their extrapolation. We argue that Alfred North Whitehead’s concept of ‘abstraction’ can provide an alternative appreciation of some key aspects of the processes of knowledge-production of RCTs to enable a recasting of the problem of generalization. By proposing that generalization depends on relevant abstractions, we direct attention to the situated forms of care that this calls for. After showing the conceptual difference that the process of abstraction makes for understanding and extrapolating the situated nature of a research finding, we offer an interpretation of possible forms of care at work in efforts to devise Ebola adaptive trials. The example is offered as one possible basis for a reformulation of the logic of generalization.     

Promoting social skill use and generalization in children with autism spectrum disorder

Social skills training is a frequently utilized intervention for addressing social deficits of children with autism spectrum disorder (ASD). The current study investigated the effects of the Superheroes Social Skills program, a social skills training program consisting of multiple evidence-based practices, in promoting accurate demonstration of target social skills in both a training and generalized setting. Three children with ASD between the ages of 10 and 14 attended 10 social skills training sessions over five weeks, with social skills lessons targeting participation, conversation, perspective taking, and problem solving skills. A multiple probe design across behaviors and replicated across participants was utilized to evaluate accurate demonstration of target social skills. Results suggest improvement in skill accuracy in both the training and generalization settings. Indirect secondary measures of participant social functioning and parental stress were also collected and suggest improvements associated with social skills training.     

Inoculation Against Falls: Rapid Adaptation by Young and Older Adults to Slips During Daily Activities

Pai Y-C, Bhatt T, Wang E, Espy D, Pavol MJ. Inoculation against falls: rapid adaptation by young and older adults to slips during daily activities.

Objective

To determine whether aging diminishes one's ability to rapidly learn to resist falls on repeated-slip exposure across different activities of daily living.

Design

Quasi-experimental controlled trial.

Setting

Two university-based research laboratories.

Participants

Young (n=35) and older (n=38) adults underwent slips during walking. Young (n=60) and older (n=41) adults underwent slips during a sit-to-stand task. All (N=174) were healthy and community dwelling.

Intervention

Low-friction platforms induced unannounced blocks of 2 to 8 repeated slips interspersed with blocks of 3 to 5 nonslip trials during the designated task.

Main Outcome Measures

The incidence of falls and balance loss. Dynamic stability (based on center of mass position and velocity) and limb support (based on hip height) 300ms after slip onset.

Results

Under strictly controlled, identical low-friction conditions, all participants experienced balance loss, but older adults were over twice as likely as young to fall on the first, unannounced, novel slip in both tasks. Independent of age or task, participants adapted to avoid falls and balance loss, with most adaptation occurring in early trials. By the fifth slip, the incidence of falls and balance loss was less than 5% and 15%, respectively, regardless of age or task. Reductions in falls and balance loss for each task were accomplished through improved control of stability and limb support in both age groups. A rapidly reversible age- and task-dependent waning of motor learning occurred after a block of nonslip trials. Adaptation to walk slips reached a steady state in the second slip block regardless of age.

Conclusions

The ability to rapidly acquire fall-resisting skills on repeated-slip exposure remains largely intact at older ages and across functional activities. Thus, repeated-slip exposure might be broadly effective in inoculating older adults against falls.     

Functional analysis and treatment of problem behavior exhibited by children with fragile X syndrome

The efficacy of function-based interventions for the treatment of severe problem behavior exhibited by individuals with intellectual and developmental disabilities (IDD) is well established. However, few studies have reported on behavioral interventions in fragile X syndrome (FXS) specifically. The present study is a consecutive case-series analysis that reports on functional analysis and treatment of problem behavior of nine children with FXS. Assessment findings were consistent with previous research indicating that among individuals with FXS, problem behavior is more commonly maintained by escape from demands and access to tangible items, relative to the broader population of individuals with IDD. Functional analysis-based behavioral interventions resulted in a mean reduction in problem behavior of 95.2% across the nine participants. Additionally, generalization of treatment effects from controlled clinical settings to home, school, and community was demonstrated. The current findings suggest that function-based behavioral interventions shown to be effective with the broader population of individuals with IDD are also effective for individuals with FXS. Our results in combination with those of previous studies describing functional analysis outcomes provide additional evidence for a unique functional behavioral phenotype for severe problem behavior in individuals with FXS. Implications of study findings for early intervention and prevention of problem behavior in children with FXS are discussed.     

Depression impairs learning,whereas the selective serotonin reuptake inhibitor,paroxetine, impairs generalization in patients with major depressive disorder

To better understand how medication status and task demands affect cognition in major depressive disorder (MDD), we evaluated medication-naïve patients with MDD, medicated patients with MDD receiving the selective serotonin reuptake inhibitors (SSRI) paroxetine, and healthy controls. All three groups were administered a computer-based cognitive task with two phases, an initial phase in which a sequence is learned through reward-based feedback (which our prior studies suggest is striatal-dependent), followed by a generalization phase that involves a change in the context where learned rules are to be applied (which our prior studies suggest is hippocampal-region dependent). Medication-naïve MDD patients were slow to learn the initial sequence but were normal on subsequent generalization of that learning. In contrast, medicated patients learned the initial sequence normally, but were impaired at the generalization phase. We argue that these data suggest (i) an MDD-related impairment in striatal-dependent sequence learning which can be remediated by SSRIs and (ii) an SSRI-induced impairment in hippocampal-dependent generalization of past learning to novel contexts, not otherwise seen in the medication-naïve MDD group. Thus, SSRIs might have a beneficial effect on striatal function required for sequence learning, but a detrimental effect on the hippocampus and other medial temporal lobe structures is critical for generalization.     

A local Vapnik–Chervonenkis complexity

We define in this work a new localized version of a Vapnik–Chervonenkis (VC) complexity, namely the Local VC-Entropy, and, building on this new complexity, we derive a new generalization bound for binary classifiers. The Local VC-Entropy-based bound improves on the original Vapnik’s results because it is able to discard those functions that, most likely, will not be selected during the learning phase. The result is achieved by applying the localization principle to the original global complexity measure, in the same spirit of the Local Rademacher Complexity. By exploiting and improving a recently developed geometrical framework, we show that it is also possible to relate the Local VC-Entropy to the Local Rademacher Complexity by finding an admissible range for one given the other. In addition, the Local VC-Entropy allows one to reduce the computational requirements that arise when dealing with the Local Rademacher Complexity in binary classification problems.     

Explorations of the mean field theory learning algorithm

The mean field theory (MFT) learning algorithm is elaborated and explored with respect to a variety of tasks. MFT is benchmarked against the back-propagation learning algorithm (BP) on two different feature recognition problems: two-dimensional mirror symmetry and multidimensional statistical pattern classification. We find that while the two algorithms are very similar with respect to generalization properties, MFT normally requires a substantially smaller number of training epochs than BP. Since the MFT model is bidirectional, rather than feed-forward, its use can be extended naturally from purely functional mappings to a content addressable memory. A network with N visible and N hidden units can store up to approximately 4N patterns with good content-addressability. We stress an implementational advantage for MFT: it is natural for VLSI circuitry.     

A Review of Challenges in the Use of fMRI for Disease Classification / Characterization and A Projection Pursuit Application from A Multi-site fMRI Schizophrenia Study

Functional magnetic resonance imaging (fMRI) is a fairly new technique that has the potential to characterize and classify brain disorders such as schizophrenia. It has the possibility of playing a crucial role in designing objective prognostic/diagnostic tools, but also presents numerous challenges to analysis and interpretation. Classification provides results for individual subjects, rather than results related to group differences. This is a more complicated endeavor that must be approached more carefully and efficient methods should be developed to draw generalized and valid conclusions out of high dimensional data with a limited number of subjects, especially for heterogeneous disorders whose pathophysiology is unknown. Numerous research efforts have been reported in the field using fMRI activation of schizophrenia patients and healthy controls. However, the results are usually not generalizable to larger data sets and require careful definition of the techniques used both in designing algorithms and reporting prediction accuracies. In this review paper, we survey a number of previous reports and also identify possible biases (cross-validation, class size, e.g.) in class comparison/prediction problems. Some suggestions to improve the effectiveness of the presentation of the prediction accuracy results are provided. We also present our own results using a projection pursuit algorithm followed by an application of independent component analysis proposed in an earlier study. We classify schizophrenia versus healthy controls using fMRI data of 155 subjects from two sites obtained during three different tasks. The results are compared in order to investigate the effectiveness of each task and differences between patients with schizophrenia and healthy controls are investigated.     

Contingency Learning Deficits and Generalization in Chronic Unilateral Hand Pain Patients

Contingency learning, in particular the formation of danger beliefs, underpins conditioned fear and avoidance behavior, yet equally important is the formation of safety beliefs. That is, when threat beliefs and accompanying fear/avoidance spread to technically safe cues, it might cause disability. Indeed, such overgeneralization has been advanced as a transdiagnostic pathologic marker, but it has not been investigated in chronic pain. Using a novel hand pain scenario contingency learning task, we tested the hypotheses that chronic hand pain patients demonstrate less differential pain expectancy judgments because of poor safety learning and demonstrate broader generalization gradients than healthy controls. Participants viewed digitized 3-dimensional hands in different postures presented in random order (conditioned stimulus [CS]) and rated the likelihood that a fictive patient would feel pain when moving the hand into that posture. Subsequently, the outcome (pain/no pain) was presented on the screen. One hand posture was followed by pain (CS+), another was not (CS−). Generalization was tested using novel hand postures (generalization stimuli) that varied in how similar they were to the original conditioned stimuli. Patients, but not healthy controls, demonstrated a contingency learning deficit determined by impaired safety learning, but not by exaggerated pain expectancy toward the CS+. Patients showed flatter, asymmetric generalization gradients than the healthy controls did, with higher pain expectancy for novel postures that were more similar to the original CS−. The results clearly uphold our hypotheses and suggest that contingency learning deficits might be important in the development and maintenance of the chronic pain–related disability.