Ee 102b Signal Processing And Linear Systems Ii

[Mandy Geise]

Appliedlinear algebra and linear dynamical systems with applications to circuits, signal processing, communications, and control systems. Topics: least-squares approximations of over-determined equations, and least-norm solutions of underdetermined equations. Symmetric matrices, matrix norm, and singular-value decomposition. Eigenvalues, left and right eigenvectors, with dynamical interpretation.

Matrix exponential, stability, and asymptotic behavior. Multi-input/multi-output systems, impulse, and step matrices; convolution and transfer-matrix descriptions. Control, reachability, and state transfer; observability and least-squares state estimation.

Hosted through the Summer Academic Resource Center (SARC), we offer a variety of educational workshops to complement your academic pursuits. Presenters varying from Stanford affiliates, community partners, and graduate tutors will cover a wide range of topics from academic skills to career exploration. Past workshops included college preparedness, software exploration, and building your professional network.

Where will your studies take you?

At educations.com, we believe that students who study abroad become the next generation of globally-minded adventurers and leaders - and we want more of you to do it! Every year, our search engine helps over 8 million students connect with some of the best universities and schools around the world.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Fronto-striatal circuits have been implicated in cognitive control of behavioral output for social and appetitive rewards. The functional diversity of prefrontal cortical populations is strongly dependent on their synaptic targets, with control of motor output mediated by connectivity to dorsal striatum. Despite evidence for functional diversity along the anterior-posterior striatal axis, it is unclear how distinct fronto-striatal sub-circuits support value-based choice. Here we found segregated prefrontal populations defined by anterior/posterior dorsomedial striatal target. During a feedback-based 2-alternative choice task, single-photon imaging revealed circuit-specific representations of task-relevant information with prelimbic neurons targeting anterior DMS (PL::A-DMS) robustly modulated during choices and negative outcomes, while prelimbic neurons targeting posterior DMS (PL::P-DMS) encoded internal representations of value and positive outcomes contingent on prior choice. Consistent with this distributed coding, optogenetic inhibition of PL::A-DMS circuits strongly impacted choice monitoring and responses to negative outcomes while inhibition of PL::P-DMS impaired task engagement and strategies following positive outcomes. Together our data uncover PL populations engaged in distributed processing for value-based choice.

Seminal studies in rat provided the first evidence of functional segregation along the striatal A-P axis, with posterior dorsomedial striatum (P-DMS) lesions disrupting both the initial acquisition and post-training execution of instrumental conditioning, in particular modulation of response according to action-outcome association8,9. In contrast, the importance of the anterior dorsomedial striatum (A-DMS) in goal-directed choice remained uncertain, with opposing results for pharmacological inactivation and excitotoxic lesions8,9,12. Optogenetic manipulations of specific spiny projection neuron subtypes within the A-DMS have implicated this subregion in supporting flexible responses during reversal learning13, consistent with pharmacological manipulations of anterior caudate in marmosets14. In contrast, the anterior dorsolateral striatum (DLS) supports a protein synthesis-dependent consolidation of newly learned actions15. Finally, a growing body of evidence has implicated the rodent striatal tail, the most caudal subregion, in behavioral responses to aversive stimuli and psychostimulants16,17,18.

The prefrontal cortex exerts cognitive control over mammalian behavior via extensive afferent integration and widespread downstream connectivity19,20. Analysis of prefrontal populations accounting for downstream synaptic targets has revealed pathway-specific functional differences for prefrontal control of social-spatial rewards21, reward anticipation22, and choice directions23. The prelimbic region of the prefrontal cortex has been hypothesized to support goal-directed action by encoding short-term memories necessary for subsequent action-outcome associations in dorsal striatum24. Specific targeting of prelimbic-striatal pathways has extended this view, demonstrating persistent neural coding of value essential for choice behavior25 and the mediation of response inhibition during tasks requiring sustained attention26. Finally, DREADD-mediated inhibition of PL neurons projecting to either anterior or posterior striatal subregions has uncovered involvement in instrumental learning6,10.

Here we systematically explore the function of PL pathways projecting along the A-P striatal axis in male mice via integration of mono- and di-synaptic viral circuit tracing, local circuit connectivity measures, single neuron calcium imaging, statistical modeling of neural coding properties, and target-specific optogenetic manipulations. Retrograde tracing from A/P-DMS subregions revealed non-overlapping PL populations, which exhibited unique encoding of behavioral variables over multiple time scales essential for shaping efficient action selection and execution. Target- and temporally- specific optogenetic manipulations confirmed the functional divergence of these fronto-striatal pathways, with PL::A-DMS pathways supporting choice monitoring and responding to negative outcomes and PL::P-DMS pathways supporting task motivation and responding to positive outcomes. Together, our results provide insight into the distributed nature of fronto-striatal pathways for decision making.

To explore whether distinct afferent connectivity could explain previously described differences in DMS function along the anterior-posterior axis8,27, we injected two distinct Alexa-conjugated Cholera toxin subunit-B retrograde tracers into A-DMS and P-DMS (Fig. S1a). Excepting the amygdala, most afferent projection regions targeted both DMS striatal areas with non-overlapping neuronal populations (Fig. S1j). To better understand the functional implications of this unique circuit architecture, we focused on prefrontal cortical areas, particularly prelimbic cortex (PL), which despite a bias towards A-DMS, targeted both DMS compartments (Fig. S1j). To confirm that synaptic inputs from PL were spread along the full anterior-posterior extent of DMS, we injected a mix of AAV5-CamKII::GFP-Cre and AAVdj-EF1a::Flex-Synaptophysin-mRuby virus into PL (Fig. 1a, b).

Beyond postsynaptic connectivity, another source of circuit diversity lies in the inputs that neurons receive. To investigate this, we examined second-order connectomes for PL neurons defined by A/P-DMS subregion by injecting retroAAV2-EF1a::3xFLAG-Cre into either A- or P-DMS subregions and a mixture of AAV-DJ-CAG::FLEX-TVA-mCherry and AAV-DJ-CAG::DIO-RVG into PL cortex (Fig. 2d). Subsequent PL injection of EnvA-RV-EGFP permitted single synapse tracing specifically from PL neurons that projected to either DMS subregion (2nd order inputs). Consistent with these fronto-striatal circuits being embedded in the same local microcircuit, we observed multiple afferent populations with similar targeting of each PL circuit, including dorsal anterior cingulate cortex (dACC) and both associative and ventral motor thalamic nuclei (Fig. 2e, f). Surprisingly though, we also noted pathway-specific distinctions in second order afferent connections, with strong PL::P-DMS biases from secondary motor cortex (M2) and significant PL::A-DMS biases from ventral anterior cingulate cortex, retrosplenial cortex and orbitofrontal cortex. Together, these data show that distinct PL populations determined by A/P-DMS target have distinct striatal synaptic connectivity and biases in their afferent drivers.

To broadly assess pathway-specific PL tuning to events preceding outcome feedback, we grouped the cue, initiation, and choice predictors. We found that PL::A-DMS circuits were more strongly tuned to these pre-outcome events (Fig. 5a,b), with the majority of modulation driven by choice-associated tuning (Fig. 5c). In our encoding model, choice-associated activity may reflect components of action selection (decision process, motor command, ongoing motor kinematics) or choice evaluation (predictive or efference signals). We found that PL::A-DMS encoded both ipsilateral and contralateral choices (Fig. S6a, b) in a near exclusive manner (Fig. 5d). Despite the PL::A-DMS pathway bias for generally encoding movement kinematics (Fig. S6h), we found that the majority of choice modulated neurons did not strongly encode head velocity (Fig. 5e and Fig. S6i), nor the specific turning actions required to enter ports (Fig. S6j).

Thus far, our data highlight a unique PL-striatal architecture defined by A-P striatal target that encodes complementary aspects of a value-based choice task. Our neural coding analysis makes several predictions about pathway-specific behavioral functions: 1. PL::A-DMS choice activity may shape current choice execution or instead provide an action-monitoring/expectation signal; 2. the persistent choice x positive outcome activity in PL::P-DMS could be used to drive positive reinforcement behavior; 3. PL::A-DMS negative outcome modulated neurons could be used to regulate choice strategies following negative outcome; 4. PL::P-DMS neurons encode temporally integrated signals for local reward rate and action value that may drive task engagement.

3a8082e126