慢性疲劳综合征研究揭示能力限制而非意愿不足

英文原文

In a recent, high-profile study of post-infectious myalgic encephalomyelitis/chronic fatigue syndrome (PI-ME/CFS), Walitt et al. (2024) assessed the performance of patients and healthy volunteers on the Effort-Expenditure for Rewards Task (EEfRT), among a host of other measures. The EEfRT is a widely used behavioral index of reward motivation and effort-based decision-making that requires repeatedly choosing between an easy task and a hard task, each involving rapid, repetitive button-pressing (Treadway et al., 2009). Walitt et al.'s study—the first to investigate effort-based decision-making in PI-ME/CFS—found that patients were less likely to choose the hard task than healthy volunteers. The authors interpreted this difference as evidence of altered “effort preference,” which they defined as “how much effort a person subjectively wants to exert” (p. 9). Walitt et al. concluded that “effort preference, not fatigue, is the defining motor behavior of this illness” (p. 10). Here we interrogate this conclusion. Were PI-ME/CFS patients less likely to choose the hard task because they wanted to exert less effort, consciously or otherwise? Or were they less able to complete the hard task, and thus chose it less often? We argue that the data support the latter interpretation. For the EEfRT to yield interpretable results, participants' choices between the easy and hard tasks must be decoupled from their ability to complete these tasks. As the developers of the measure cautioned, “an important requirement for the EEfRT is that it measure individual differences in motivation for rewards, rather than individual differences in ability or fatigue” (Treadway et al., 2009, p. 4). Several techniques have been established to satisfy this requirement. In an initial validation study with a healthy student sample, Treadway et al. (2009) ruled out differences in both ability and within-task fatigue through two manipulation checks: confirmation of ceiling-level trial completion rates for all participants (96–100%) and the inclusion of trial number in statistical models. Subsequent studies have prospectively assessed and statistically controlled for motor skills, which have been shown to influence effort-based decision-making (Ohmann et al., 2020, 2022). In studies of schizophrenia, where patients have exhibited lower maximum button press rates than controls, the required number of button presses is often calibrated to individual ability levels (e.g., Cooper et al., 2019; Fervaha et al., 2013; Le et al., 2023; Reddy et al., 2015). This method helps fulfill the prerequisite, affirmed by the EEfRT developers, that “all subjects [are] readily able to complete both the hard and easy tasks throughout the experiment” (Treadway et al., 2009, p. 4; emphasis added). Despite the importance of ensuring that decision-making in the EEfRT cannot be explained by ability or fatigue, Walitt et al. (2024) only ruled out fatigue.^1 In their study, 15 PI-ME/CFS patients and 16 healthy controls chose between the easy and hard tasks an average of 46 times during the 15-minute testing session. The easy task required 30 button presses in 7 seconds with the dominant index finger, while the hard task required 98 button presses in 21 seconds with the non-dominant little finger.^2 Whereas the EEfRT is typically regarded as a measure of reward motivation, effort-based decision-making, or effort allocation (Cooper et al., 2019; Fervaha et al., 2013; Le et al., 2023; Ohmann et al., 2020, 2022; Reddy et al., 2015; Treadway et al., 2009), Walitt et al. characterized it as assessing the new construct of “effort preference, the decision to avoid the harder task when decision-making is unsupervised and reward values and probabilities of receiving a reward are standardized” (p. 2). Effort preference was operationalized as the Proportion of Hard Task Choices (PHTC). Compared to the PI-ME/CFS patients, the controls were 1.65 times more likely to choose the hard task (PHTC) and, upon choosing it, were 27.23 times more likely to successfully complete it. Walitt et al. interpreted the difference in PHTC as indicating that people with PI-ME/CFS prefer to avoid effort, ignoring the substantially larger—and more fundamental—difference in the ability of patients and controls to complete the hard task successfully. Our own analysis of Walitt et al.'s (2024) data, accessed via mapMECFS (Mathur et al., 2021), confirmed the large difference in ability between the groups (analysis code available at https://osf.io/vqzca/). PI-ME/CFS patients were only able to complete an average of 65% of the hard tasks they chose (SD = 37%), compared to 96% (SD = 8%) for controls, two-sided Mann-Whitney U: p = 0.01, r = 0.45 (non-parametric analyses were used to account for extreme negative skew in completion rates). This constitutes a failure of Treadway et al.'s (2009) manipulation check for differences in ability. Moreover, as shown in Figure 1, seven of the 15 PI-ME/CFS patients had lower hard task completion rates than any control, successfully completing only 30% of the hard tasks they chose (SD = 23%). These results show that the hard task was simply too hard for many PI-ME/CFS patients. By contrast, both groups completed the easy task at near-ceiling rates (PI-ME/CFS: M = 98%, SD = 4%; controls: M = 99%, SD = 2%), two-sided Mann-Whitney U: p = 0.63. The stark contrast in performance on the two tasks suggests that, for the PI-ME/CFS patients, choosing the hard task over the easy one (PHTC) may have been more a matter of ability than preference. Consistent with this possibility, we found that PHTC was positively correlated with the proportion of hard tasks completed successfully, r_s(29) = 0.38, p = 0.03 (see Figure 1). These results suggest that PHTC was confounded with ability in Walitt et al.'s (2024) study: participants who had more difficulty completing the hard task chose it less often. Because difficulty with the hard task disproportionately affected the PI-ME/CFS group, the large difference in hard task completion rates could explain the comparatively small difference in PHTC between groups. Therefore, we interpret Walitt et al.'s data as showing that people with PI-ME/CFS were less able to execute the EEfRT's hard task, rather than unwilling to expend effort. In sum, Walitt et al.'s (2024) data provide no evidence of altered effort preference in PI-ME/CFS patients, who lacked the physical ability to consistently execute the task assessing it. Conclusions about effort preference are unwarranted when group differences in ability could account for disparities in task performance. To decouple what patients are willing to do from what they are able to do, future research in ME/CFS should calibrate measures of effort-based decision-making to the ability of individual patients. The amount of effort a person wants to exert on a task is irrelevant if they are unable to exert it.

中文翻译

在最近一项关于感染后肌痛性脑脊髓炎/慢性疲劳综合征（PI-ME/CFS）的高调研究中，Walitt等人（2024年）评估了患者和健康志愿者在奖励任务努力支出（EEfRT）以及其他一系列测量中的表现。EEfRT是一种广泛使用的奖励动机和基于努力的决策行为指标，要求参与者在简单任务和困难任务之间反复选择，每个任务都涉及快速、重复的按键操作（Treadway等人，2009年）。Walitt等人的研究——首次调查PI-ME/CFS中基于努力的决策——发现患者选择困难任务的可能性低于健康志愿者。作者将这种差异解释为“努力偏好”改变的证据，他们将其定义为“一个人主观上想要付出多少努力”（第9页）。Walitt等人得出结论：“努力偏好，而非疲劳，是这种疾病的决定性运动行为”（第10页）。在此，我们质疑这一结论。PI-ME/CFS患者选择困难任务的可能性较低，是因为他们有意或无意地想要付出更少的努力吗？还是因为他们完成困难任务的能力较差，因此选择它的频率较低？我们认为数据支持后一种解释。 为了使EEfRT产生可解释的结果，参与者在简单任务和困难任务之间的选择必须与完成这些任务的能力脱钩。正如该测量的开发者所警告的，“EEfRT的一个重要要求是测量奖励动机的个体差异，而不是能力或疲劳的个体差异”（Treadway等人，2009年，第4页）。已经建立了多种技术来满足这一要求。在一项针对健康学生样本的初步验证研究中，Treadway等人（2009年）通过两个操作检查排除了能力和任务内疲劳的差异：确认所有参与者的试验完成率达到天花板水平（96-100%）以及在统计模型中包含试验编号。后续研究前瞻性评估并统计控制了运动技能，这些技能已被证明会影响基于努力的决策（Ohmann等人，2020年，2022年）。在精神分裂症的研究中，患者表现出比对照组更低的最高按键率，所需按键次数通常根据个体能力水平进行校准（例如，Cooper等人，2019年；Fervaha等人，2013年；Le等人，2023年；Reddy等人，2015年）。这种方法有助于满足EEfRT开发者确认的前提条件，即“所有受试者在整个实验过程中都能轻松完成困难和简单任务”（Treadway等人，2009年，第4页；强调添加）。 尽管确保EEfRT中的决策不能由能力或疲劳解释至关重要，但Walitt等人（2024年）仅排除了疲劳。^1 在他们的研究中，15名PI-ME/CFS患者和16名健康对照组在15分钟的测试会话中平均在简单任务和困难任务之间选择了46次。简单任务要求用优势食指在7秒内按键30次，而困难任务要求用非优势小指在21秒内按键98次。^2 尽管EEfRT通常被视为奖励动机、基于努力的决策或努力分配的测量（Cooper等人，2019年；Fervaha等人，2013年；Le等人，2023年；Ohmann等人，2020年，2022年；Reddy等人，2015年；Treadway等人，2009年），但Walitt等人将其描述为评估新的构念“努力偏好，即在决策无监督且奖励价值和获得奖励的概率标准化时避免更困难任务的决定”（第2页）。努力偏好被操作化为困难任务选择比例（PHTC）。与PI-ME/CFS患者相比，对照组选择困难任务（PHTC）的可能性高出1.65倍，并且在选择后成功完成的可能性高出27.23倍。Walitt等人将PHTC的差异解释为表明PI-ME/CFS患者偏好避免努力，忽略了患者和对照组成功完成困难任务能力的更大且更根本的差异。 我们自己对Walitt等人（2024年）数据的分析（通过mapMECFS访问；Mathur等人，2021年）确认了两组之间能力的巨大差异（分析代码可在https://osf.io/vqzca/获取）。PI-ME/CFS患者仅能完成平均65%他们选择的困难任务（SD = 37%），而对照组为96%（SD = 8%），双侧Mann-Whitney U检验：p = 0.01，r = 0.45（使用非参数分析以考虑完成率的极端负偏态）。这构成了Treadway等人（2009年）对能力差异操作检查的失败。此外，如图1所示，15名PI-ME/CFS患者中有7名的困难任务完成率低于任何对照组，仅成功完成30%他们选择的困难任务（SD = 23%）。这些结果表明，困难任务对许多PI-ME/CFS患者来说根本太难了。相比之下，两组在简单任务上的完成率都接近天花板水平（PI-ME/CFS：M = 98%，SD = 4%；对照组：M = 99%，SD = 2%），双侧Mann-Whitney U检验：p = 0.63。 两项任务表现的鲜明对比表明，对于PI-ME/CFS患者来说，选择困难任务而非简单任务（PHTC）可能更多是能力问题而非偏好问题。与这种可能性一致，我们发现PHTC与成功完成困难任务的比例呈正相关，r_s(29) = 0.38，p = 0.03（见图1）。这些结果表明，在Walitt等人（2024年）的研究中，PHTC与能力混淆：完成困难任务难度更大的参与者选择它的频率更低。由于困难任务的难度不成比例地影响了PI-ME/CFS组，困难任务完成率的巨大差异可以解释两组之间PHTC的相对较小差异。因此，我们将Walitt等人的数据解释为表明PI-ME/CFS患者执行EEfRT困难任务的能力较低，而非不愿意付出努力。 总之，Walitt等人（2024年）的数据没有提供PI-ME/CFS患者努力偏好改变的证据，这些患者缺乏持续执行评估任务的身体能力。当能力差异可以解释任务表现差异时，关于努力偏好的结论是不合理的。为了将患者愿意做的事情与他们能够做的事情脱钩，未来ME/CFS研究应校准基于努力的决策测量以适应个体患者的能力。如果一个人无法付出努力，那么他们想要在任务上付出多少努力是无关紧要的。

文章概要

本文针对关键词“Exception-finding in goal setting for managing chronic fatigue syndrome”，批判性地分析了Walitt等人（2024年）关于PI-ME/CFS患者努力偏好的研究。文章指出，该研究使用EEfRT任务时未充分校准患者的能力限制，导致将患者选择困难任务较少归因于“努力偏好”而非实际能力不足。通过数据分析，作者发现PI-ME/CFS患者完成困难任务的成功率显著低于对照组（65% vs 96%），且困难任务完成率与选择比例正相关，表明能力差异是主要因素。文章强调，在目标设定中识别例外（如能力限制）至关重要，未来研究应个性化校准测量工具，以准确区分意愿与能力，从而更有效地管理慢性疲劳综合征。

高德明老师的评价