Gordon Banks

Radiologic automated diagnosis (RAD)

RAD is a program currently being developed to interpret neuroimages. Given the clinical information usually available on the imaging request, RAD will analyze the scan directly from the data generated by the scanning machine to produce a differential diagnostic list explaining any lesions it discovers. RAD uses a computerized three-dimensional stereotaxic atlas of the nervous system as a model of normal structures in the analysis of scans.

Symbolic coordinate anatomy for neurology (SCAN).

Our laboratory is responsible for development and maintenance of CADUCEUS (Formerly INTERNIST-1), a computer program to provide expert advice to the clinician in the realm of medical diagnosis. In order to extend its success in internal medicine into neurology, it was necessary to provide a paradigm for neuroanatomic reasoning. In our anatomic knowledge base (SCAN), the nervous system is partitioned into a hierarchical set of nested cubes. In the computer memory, a symbol for each cube has attached to it lists of structures that are wholly or partially within the cube. The structures may be simple (e.g.,right locus ceruleus) or complex (e.g.,midbrain). Lists of cubes belonging to particular vascular territories as well as systems (e.g.,visual) are also maintained. Thus, computation of anatomic localization from a given symptom or finding is facilitated as well as characterization of consequences of vascular lesions or neighborhood effects from mass lesions.

INKBLOT: A neurological diagnostic decision support system integrating causal and anatomical knowledge

As an initial step in the diagnostic process, human neurologists often use anatomical localization to constrain the set of diagnostic hypotheses deserving further consideration. We describe an automated system, INKBLOT-1, which uses anatomical localization in much the same way as human neurologists. Given a set of manifestations, INKBLOT-1 generates a set of hypothetical localizations relative to a coordinate system of nested cubes and then uses these localization(s) to explain the manifestations. We trace the reasoning mechanism utilized by INKBLOT-1 for a particular set of symptoms and show how INKBLOT-1 is able to generate novel hypotheses that explain the observed manifestations. In doing this, INKBLOT-1 demonstrates capabilities not demonstrated by previously described systems.

An Information-Based Bayesian Approach to History Taking

Effective history-taking systems need to dynamically reduce the number of questions to ask. This can be done either categorically or probabilistically, by exploiting previous patients answers. In this paper, we propose a probabilistic information-based history-taking strategy that combines synergistically two information-content measures for reducing the number of questions asked. We have applied this strategy to an existing history-taking system and some preliminary results seem to confirm our initial intuitions.

Speech motor control disorder after HIV infection

We examined the clinical characteristics of six right-handed patients who developed speech motor control disorders after human immunodeficiency virus (HIV) infection. They exhibited an ataxic dysarthria, characterized by irregular articulatory breakdowns in consonant and vowel timing; were slow in timed decision-making tasks; and had impaired procedural learning. Other aspects of the neurologic examination revealed signs of diffuse CNS involvement including action-intention tremors, ataxic gait, and release signs. None developed HIV-associated dementia during 1 year of follow-up. Motor speech control disorder appears to be related to a cerebellar dysfunction.

Methodological considerations in estimating speed of cognitive operations.

Individuals infected with Human Immunodeficiency Virus (HIV) and having cognitive impairment have been described as having slow mentation. Data supporting this proposition come from a variety of sources, including Sternbergs (1966) item recognition memory task. The procedure nominally provides an index of speed of mental operations, independent from input/output demands. However, since the original use of this procedure in the 1960s, advances in cognitive psychology have revealed many of its limitations. The purpose of the present study was to examine the psychometric characteristics of this task. Each participant performed the Sternberg item recognition task twice, 6 mo apart. The stability of the estimate of the slope of regression equations and for zero intercept ranged from excellent (r = .87) to poor (r = .30), and the data from many individual subjects could not be reliably modelled using multiple linear regression techniques. These data, as well as those from previous research, demonstrate the limited practical use of this task in clinical samples. Furthermore, as cognitive psychological theory has advanced in the past 30 yr, the conceptual underpinnings of the procedure have essentially evaporated.

Computerized Three-Dimensional Stereotactic Atlases

The goal of functional stereotactic surgery is to stimulate or ablate selected targets in the brain. The usual targets are subcortical nuclei and white matter tracts. To accomplish this goal, the three-dimensional coordinates of the target sites must be precisely determined. The traditional tool for this purpose has been the stereotactic atlas.

Development of Subtle Neurological Signs after Systemic Illness in HIV-infected Individuals

Thirty human immunodeficiency virus (HIV) infected individuals entered a longitudinal study without signs of dysfunction of the central nervous system (CNS). Nine of these individuals developed a systemic illness between study visits, and 7 of these 9 patients (78%) had neurological signs at the next examination (e.g., action-intention tremors, abnormal gait, release signs, abnormal deep tendon reflexes). Only 2/21 (9.5%) of the subjects who did not develop systemic illness showed such signs. These data are consistent with the hypothesis that other factors (e.g. cytokines) as well as the HIV may cause subtle CNS dysfunction.

The michael reese hospital stroke registry a microcomputer-implemented data base

Abstract A properly implemented patient registry can greatly facilitate clinical research, serving to establish or negate commonly accepted manifestations of disease which are based on anecdotal evidence and experience, as well as uncover previously unrecognized correlations. It can provide lists of patients with any desired profile for use in new clinical studies. With the advent of inexpensive microcomputers, many clinicians now can effectively use patient registries. We have constructed a program to accumulate a registry of stroke patients. It stores 328 items of information about each patient, including history, physical findings, laboratory and radiological data and pathological findings. A search program will allow retrieval using Boolean combinatorial search on any of the keys. The program is implemented on an Apple II computer. At present, more than 500 patients have been entered into the registry.