On the meaning and measurement of affective instability: clues from chaos theory

Abstract

Background: “Affective instability” describes affective shifts occurring over hours to days that are associated with clinically significant impairment or distress. Since patients with this condition are clinically “chaotic,” we undertook to study affective instability relative to normal affective variability using the tools of chaos theory.

Methods: Patients and controls generated time series data over 90 days using a visual analog mood scale to capture daily affective means and extremes. The series were analyzed using the Mean Squared Successive Difference (MSSD), Power Spectral Density (PSD), and Fractal Dimension (FD).

Results: Patients demonstrated substantially more variability than controls on the MSSD, but less complexity as measured by the FD. The PSD revealed that power varied with frequency (f) in a 1f^α relationship, wherein the α for patients was double that for controls.

Conclusions: Despite the “chaotic” clinical presentation of affective instability patients, affective instability itself was found to be less complex from a chaos-theoretic perspective than normal affective variability. Of particular interest is the α ratio of order 2 between patients and controls seen in both our study and a similar but much longer study of mood in rapid-cycling bipolar disorder; an observation suggesting that pathological affect may be distinguishable from normal affective variability by a scale-invariant parameter.

Introduction

Affective instability is a term of great clinical relevance referring to affective shifts occurring over hours to days that are associated with clinically significant impairment or distress. These shifts are notable for their abrupt and irregular nature. The irregularity of affect over time is both a pervasive characteristic of affective disorders and a disruptive influence on attempts at classification, as noted by Kraepelin in his Manic-Depressive Insanity and Paranoia(Kraepelin 1921): “I think that I am convinced that … effort at classification must of necessity wreck on the irregularity of the disease. The kind and duration of the attacks and the intervals by no means remain the same in the individual case but may frequently change, so that the case must be reckoned always to new forms” (p 139); and further on: “… no borderline at all can be drawn between the strictly periodic forms and those which run an irregular course. Of special significance for this question is the fact, that a periodicity, in some degree satisfying, exists in numerous cases only for a certain part of the course. … ” (p 188). The prevailing conceptual paradigm of organizing affective phenomena in terms of periodicity fails to capture the irregular, fluctuating features of its clinical course. Motivated by a desire to better understand these features, we sought to prospectively quantify affective instability relative to normal mood variability using the tools of chaos theory. Chaos theory has established itself in the study of highly complex phenomena, revealing that some such phenomena reflect simple but nonlinear processes, which impart measurable features that manifest across a range of time scales. This article reports on our work with affective instability that suggests that even though affective phenomena fluctuate over time scales ranging from hours to days to months to years, there may be a measure that characterizes these fluctuations relative to normal affective variability in a way that is independent of the scale of observation.