Biased credit assignment in motivational learning biases arises through prefrontal influences on striatal learning : Frontal theta/delta oscillations

<p>Parametric regressor: trial-by-trial frontal (wide ROI; t-weighted mean of channels AF3/ AF4/ AF7/ AF8/ F1/ F2/ F3/ F4/ F5/ F6/ F7/ F8/ FC1/ FC2/ FC3/ FC4/ FC5/ FC6/ FCz/ Fp1/ Fp2/ Fpz/ Fz) theta/delta oscillations (225–475 ms relative to outcome onset, 1 - 8 Hz) measured using scalp EEG. This theta response is significantly higher for negative outcomes (no rewards/ punishments) than positive outcomes (rewards/ no punishments). <br>This regressor is added on top of GLM2 regressors (featuring 8 outcome x response regressors plus standard prediction errors and the difference term to biased prediction errors), yielding GLM3B.</p><p>Motivational Go/NoGo task (Swart et al., 2017; 2018; van Nuland et al., 2020).<br></p>

### Collection description

<p>Contrast related to events around outcome receipt and learning in a Motivational Go/NoGo task (Swart et al., 2017; 2018; van Nuland et al., 2020). Note that this Go/NoGo task is equiprobable (percent Go/NoGo trials is 50:50) and thus does NOT elicit strong inhibition-related activation.</p><p>On each trial, a Win or Avoid cue appears; valence of the cue is not signaled but should be learned. Participants should respond during cue presentation. Response-dependent feedback follows after a jittered interval. <br>Each cue has only one correct action (Go-left, Go-right, or NoGo), which is followed by the desired outcome 80% of the time. For Win cues, actions can lead to rewards or neutral outcomes; for Avoid cues, actions can lead to neutral outcomes or punishments. There are eight different cues, orthogonalizing cue valence (Win versus Avoid) and required action (Go versus NoGo). Feedback is probabilistic: Correct actions to Win cues lead to rewards in 80% of cases, but neutral outcomes in 20% of cases. For Avoid cues, correct actions lead to neutral outcomes in 80% of cases, but punishments in 20% of cases. For incorrect actions, these probabilities are reversed. </p>

###Subject species

homo sapiens

###Modality

fMRI-BOLD

### Analysis level

group

### Cognitive paradigm (task)

go/no-go task

### Map type

Z

<p>参数回归模型：通过头皮脑电图（EEG）测量的试验期间前额叶（广泛区域；AF3/AF4/AF7/AF8/F1/F2/F3/F4/F5/F6/F7/F8/FC1/FC2/FC3/FC4/FC5/FC6/FCz/Fp1/Fp2/Fpz/Fz通道的t加权平均） theta/delta 脑电波（相对于结果发生的225–475 ms，频率为1 - 8 Hz）。该 theta 反应对于负面结果（无奖励/惩罚）显著高于正面结果（奖励/无惩罚）。</p><p>该回归模型在GLM2回归模型（包含8个结果 x 响应回归模型以及标准预测误差和偏差预测误差的差异项）的基础上添加，从而得到GLM3B。</p><p>动机型“进行/不进行”任务（Swart等，2017；2018；van Nuland等，2020年）。</p><p>### 数据集描述</p><p>与动机型“进行/不进行”任务中结果接收和学习相关的事件对比（Swart等，2017；2018；van Nuland等，2020年）。请注意，此“进行/不进行”任务为等概率任务（进行/不进行试验百分比各为50:50），因此不会引起强烈的抑制相关激活。</p><p>在每一轮试验中，出现“胜利”或“避免”提示；提示的效价未明确指示，但应通过学习获得。参与者应在提示呈现时做出反应。在抖动间隔后，会跟随依赖反应的反馈。</p><p>每个提示只有一个正确的动作（左进行、右进行或不进行），80%的时间内会跟随预期的结果。对于“胜利”提示，动作可能导致奖励或中性结果；对于“避免”提示，动作可能导致中性结果或惩罚。存在八个不同的提示，正交化提示效价（胜利 versus 避免）和所需动作（进行 versus 不进行）。反馈是概率性的：正确动作对于“胜利”提示在80%的情况下导致奖励，但在20%的情况下导致中性结果。对于“避免”提示，正确动作在80%的情况下导致中性结果，但在20%的情况下导致惩罚。对于错误动作，这些概率相反。</p><p>### 受试者物种</p><p>智人</p><p>### 传感模态</p><p>功能磁共振成像-BOLD</p><p>### 分析层次</p><p>组级</p><p>### 认知范式（任务）</p><p>进行/不进行任务</p><p>### 映射类型</p><p>Z</p>